Hearings
Hearing Type:
Open
Date & Time:
Wednesday, August 1, 2018 - 9:30am
Location:
Hart 216
Witnesses
Dr.
Todd
Helmus
Senior Behavioral Scientist
RAND Corporation
Ms.
Renee
DiResta
Director of Research
New Knowledge
Ms.
Laura
Rosenberger
Director
Alliance for Securing Democracy at The German Marshall Fund of the United States
Dr.
Philip
Howard
Director
Oxford Internet Institue
Full Transcript
[Senate Hearing 115-397]
[From the U.S. Government Publishing Office]
S. Hrg. 115-397
OPEN HEARING ON FOREIGN INFLUENCE
OPERATIONS' USE OF SOCIAL MEDIA PLATFORMS
(THIRD PARTY EXPERT WITNESSES)
=======================================================================
HEARING
BEFORE THE
SELECT COMMITTEE ON INTELLIGENCE
OF THE
UNITED STATES SENATE
ONE HUNDRED FIFTEENTH CONGRESS
SECOND SESSION
__________
WEDNESDAY, AUGUST 1, 2018
__________
Printed for the use of the Select Committee on Intelligence
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Available via the World Wide Web: http://www.govinfo.gov
__________
U.S. GOVERNMENT PUBLISHING OFFICE
30-957 PDF WASHINGTON : 2018
-----------------------------------------------------------------------------------
For sale by the Superintendent of Documents, U.S. Government Publishing Office,
http://bookstore.gpo.gov. For more information, contact the GPO Customer Contact Center,
U.S. Government Publishing Office. Phone 202-512-1800, or 866-512-1800 (toll-free).
E-mail, gpo@custhelp.com.
SELECT COMMITTEE ON INTELLIGENCE
[Established by S. Res. 400, 94th Cong., 2d Sess.]
RICHARD BURR, North Carolina, Chairman
MARK R. WARNER, Virginia, Vice Chairman
JAMES E. RISCH, Idaho DIANNE FEINSTEIN, California
MARCO RUBIO, Florida RON WYDEN, Oregon
SUSAN COLLINS, Maine MARTIN HEINRICH, New Mexico
ROY BLUNT, Missouri ANGUS KING, Maine
JAMES LANKFORD, Oklahoma JOE MANCHIN III, West Virginia
TOM COTTON, Arkansas KAMALA HARRIS, California
JOHN CORNYN, Texas
MITCH McCONNELL, Kentucky, Ex Officio
CHUCK SCHUMER, New York, Ex Officio
JOHN McCAIN, Arizona, Ex Officio
JACK REED, Rhode Island, Ex Officio
----------
Chris Joyner, Staff Director
Michael Casey, Minority Staff Director
Kelsey Stroud Bailey, Chief Clerk
CONTENTS
----------
AUGUST 1, 2018
OPENING STATEMENTS
Burr, Hon. Richard, Chairman, a U.S. Senator from North Carolina. 1
Warner, Mark R., Vice Chairman, a U.S. Senator from Virginia..... 3
WITNESSES
Helmus, Dr. Todd, Senior Behavioral Scientist, The Rand
Corporation.................................................... 5
Prepared statement........................................... 7
DiResta, Renee, Director of Research for New Knowledge........... 16
Prepared statement........................................... 19
Kelly, John, CEO and Founder of Graphika......................... 25
Prepared statement........................................... 27
Rosenberger, Laura, Director, Alliance for Securing Democracy,
German Marshall Fund of the United States...................... 30
Prepared statement........................................... 32
Howard, Philip, Director of the Oxford Internet Institute........ 89
Prepared statement........................................... 91
SUPPLEMENTAL MATERIAL
Responses to Questions for the Record by:
Todd Helmus.................................................. 134
Renee DiResta................................................ 142
John Kelly................................................... 148
Laura Rosenberger............................................ 150
Philip Howard................................................ 159
Charts introduced by members..................................... 163
OPEN HEARING ON FOREIGN INFLUENCE
OPERATIONS' USE OF SOCIAL MEDIA PLATFORMS
(THIRD PARTY EXPERT WITNESSES)
----------
WEDNESDAY, AUGUST 1, 2018
U.S. Senate,
Select Committee on Intelligence,
Washington, DC.
The Committee met, pursuant to notice, at 9:32 a.m. in Room
SH-216, Hart Senate Office Building, Hon. Richard Burr
(Chairman of the Committee) presiding.
Present: Senators Burr, Warner, Risch, Collins, Blunt,
Lankford, Cotton, Cornyn, Feinstein, Wyden, Heinrich, King,
Manchin, and Harris.
OPENING STATEMENT OF HON. RICHARD BURR, CHAIRMAN, A U.S.
SENATOR FROM NORTH CAROLINA
Chairman Burr. I'd like to call the hearing to order.
I'd like to welcome our witnesses today: Dr. Todd Helmus,
Senior Behavioral Scientist at the RAND Corporation; Renee
DiResta, Director of Research at New Knowledge; John Kelly, CEO
and founder of Graphika; Laura Rosenberger, Director of the
Alliance for Securing Democracy at the German Marshall Fund;
and Dr. Phil Howard, Director of the Oxford Internet Institute.
Welcome to all of you. I thank you for being here today and
for your willingness to share your expertise and insights with
this Committee and, more importantly, with the American people.
We're here to discuss a threat to the Nation that this
Committee takes every bit as seriously as terrorism, weapons of
mass destruction, espionage and regional instability. Today
we're talking about how social media platforms have enabled
foreign influence operations against the United States.
Every member of this Committee and the American people
understand what an attack on the integrity of our electoral
process means. Election interference from abroad represents an
intolerable assault on the democratic foundation this republic
was built on.
The Committee, in a bipartisan fashion, has addressed this
issue head on. In May, we released the initial findings of our
investigation into Russia's targeting of election
infrastructure during the 2016 election.
Today's hearing is an extension of that effort. But in some
ways it highlights something far more sinister, the use of our
own rights and freedoms to weaken our country from within. It's
also important that the American people know that these
activities neither began nor ended with the 2016 election. As
you can see on the one graph on display to my left, your right,
the Kremlin began testing this capability on their domestic
population several years ago, before using it against their
foes in the Near Abroad and on the United States and Western
democracies.
Even today, almost two years after the 2016 election,
foreign actors continue an aggressive and pervasive influence
campaign against the United States of America. Nothing
underscores that fact more than yesterday's announcement by
Facebook that they've identified over 30 new accounts that are
not only causing chaos in the virtual domain, but also creating
events on our streets with real Americans unknowingly
participating.
These cyber actors are using social media platforms to
spread disinformation, provoke societal conflict and undermine
public faith in democratic institutions. There does not seem to
be much debate about that.
I think it's also the case that social media isn't going
anywhere anytime soon. It's part of how we exchange ideas, we
stay connected, it binds us as a community, it gives voice to
those that are voiceless. Social media is the modern public
forum, and it's being used to divide us.
This was never about elections. It is about the integrity
of our society.
So how do you keep the good while getting rid of the bad?
That's the fundamental question in front of this Committee and
in front of the American people. And it's a complex problem
that intertwines First Amendment freedoms with corporate
responsibility, government regulation and the right of
innovators to prosper from their own work.
Sixty percent of the U.S. population uses Facebook. A
foreign power using the platform to influence how Americans
see/think about one another is as much a public policy issue as
it is a national security concern.
Crafting an elegant policy solution that's effective but
not overly burdensome demands good faith and partnership
between social media companies and this Committee. We hope to
hear from those innovators in September, because you can't
solve a problem like this by imposing a solution from 3,000
miles away. This requires a thoughtful and informed public
policy debate and this Committee is uniquely positioned to
foster that debate.
Last November, when we first welcomed the social media
companies in an open hearing, I stressed then what this debate
is and is not about. This isn't about relitigating the 2016
U.S. presidential elections. This isn't about who won or who
lost. This is about national security. This is about corporate
responsibility. And this is about the deliberate and
multifaceted manipulation of the American people by agents of a
foreign hostile government.
I thank you again for being here, for the work that you've
done. Your analytic and technical expertise is indispensable to
us getting this right. We cannot possibly formulate the right
solution without first knowing the extent of the problem.
I'm hopeful this morning that as you offer your insights
and your findings, that you'll also share your recommendations.
We can't afford ineffective half-measures, let alone nothing at
all.
While it's shocking to think that foreign actors used
social networking and communication mediums that are so central
to our lives in an effort to interfere with the core of our
democracy, what is even more troubling is that it's still
happening today. Nothing less than the integrity of our
democratic institutions, processes and ideals is at stake.
With that, I turn to the Vice Chairman.
OPENING STATEMENT OF HON. MARK R. WARNER, VICE CHAIRMAN, A U.S.
SENATOR FROM VIRGINIA
Vice Chairman Warner. Thank you, Mr. Chairman, and I also
want to welcome our witnesses today.
This Committee has invested a significant amount of time,
focus, and energy, both in public and behind closed doors, in
uncovering and exposing Russian information warfare in our own
backyard.
It is clear that our efforts have increased Americans'
understanding of what the Russians did in 2016 and how they
sought to attack us through the use of social media. It was
pressure brought by this Committee that led Facebook, Twitter
and YouTube to uncover malicious activity by the Russian-backed
Internet Research Agency. These revelations eventually resulted
in the indictments of 13 Russian individuals and three Russian
companies by the Special Counsel's Office in February of this
year.
Social media oversight has not typically been a function of
our Committee and, for that matter, any Committee. I have no
problem acknowledging that the terminology of this world--bots,
spam, click bait, API, trolls--does not always come naturally
to all of us. But thanks to bipartisan determination to
understand what happened in 2016 and a commitment to stopping
it from happening again, we have been able to accomplish a lot.
We have helped reveal the Russian playbook, we have raised
public awareness regarding the threat, and we have succeeded,
however incremental, in pressuring each of these companies to
take steps to address the problems on their platforms.
That's the good news. The bad news is that we've got a lot
more work to do. Twenty-one months after the 2016 election and
only 3 months before the 2018 elections, Russian-backed
operatives continue to infiltrate and manipulate social media
to hijack the national conversation and set Americans against
each other. They were doing it in 2016; they are still doing it
today.
That was made just evident yesterday, as the Chairman
noted, when Facebook announced the takedown of 32 new pages and
accounts that had connections to Russian-backed operations, and
those accounts had hundreds of thousands of followers.
In our previous hearings on Russian disinformation, we
outlined the Russian playbook in the 2016 elections. We
discussed how Russian operatives set up thousands of fake and
automated accounts on Facebook, Instagram, Twitter, YouTube and
others, in order to build networks of hundreds of thousands of
real Americans. These networks pushed an array of
misinformation, including stolen e-mails, state-led propaganda,
fake news and divisive content, onto the newsfeeds of as many
potentially receptive Americans as they could. And you will
note out here today from our experts that they were extremely
successful in that effort.
These active measures have two things in common: first,
they're effective; and second, they're cheap. For just pennies
on the dollar, they can wreak havoc in our society and in our
elections.
And I'm concerned that, even after 18 months of study, we
are still only scratching the surface when it comes to Russia's
information warfare. Much of the initial focus was on paid
advertisements, but it quickly became clear that these ads
represented a tiny percentage of the IRA's activity compared to
the hundreds of thousands of free Facebook and Instagram posts,
pages and groups, and millions of tweets from IRA-backed
accounts.
Today, it is becoming clearer that IRA activity represents
just a small fraction of the total Russian effort on social
media. In reality, the IRA operatives were just the incompetent
ones who made it easy to get caught. Who else is still out
there actively attacking us? Are there other troll farms? What
about the actual Russian intelligence services? I hope we'll
hear from the experts today how much further out they think
this Russian disinformation effort goes.
I'm also concerned that the United States government is not
well-positioned to detect, track or counter these types of
influence operations on social media. These types of asymmetric
attacks--which include foreign operatives appearing to be
Americans, engaging in online public discourse--almost by
design slipped between the seams of our free speech guarantees
and our legal authorities and responsibilities.
Again, I hope our witnesses will recommend ideas for better
tackling this problem while also protecting our constitutional
rights as Americans.
All the evidence this Committee has seen to date suggests
that the platform companies, namely, Facebook, Instagram,
Twitter, Google and YouTube, still have a lot of work to do.
Now, before I went into politics I spent more than 20 years in
the tech business and I have tremendous respect for these
companies and what they represent. And when they are at their
best, they are a symbol of what this country does best:
innovation, job creation, changing the world.
I've been hard on them, though, that's true. But it's
because I know they can do better to protect our democracy.
They have the creativity, expertise, resources, and
technological capability to get ahead of these malicious
actors.
That's why, as the Chairman mentioned, we'll be hosting
senior executives from Facebook, Twitter, and, yes, Google, for
a hearing on September 5th to hear the plans they have in
place, to press them to do more, and to work together to
address this challenge.
That's because it's only going to get harder. As digital
targeting continues to improve, and as new advances in
technology and artificial intelligence--one that I'm
particularly concerned on, like deep fakes--continue to spread,
the magnitude of the challenge will only grow.
I know today we'll focus on what happened in 2016 and what
is happening now, but Russian active measures have revealed a
dark underbelly of the social media ecosystem. These same tools
that spread misinformation can negatively affect other aspects
of our lives.
I think we need to start pushing ourselves beyond just
recognizing the problems and start to press actual policy ideas
forward. I'm interested in hearing some of those policy options
that might help us address broader challenges posed by the
growth and dominance of a few social media companies.
For example, does a user have the right to know if they are
interacting with a person or a bot online? Do companies have a
responsibility to ensure more transparency of how they collect,
use, and secure user data? Do users have enough control over
their own personal data?
I hope, as a panel of experts here, you can help this
Committee to lead and to begin to shape a bipartisan
responsibility to this ongoing, as the Chairman has indicated,
national security threat.
Thank you, Mr. Chairman.
Chairman Burr. I thank the Vice Chairman.
Before I move to the testimony from our witnesses, some
Committee housekeeping. After testimony, members will be
recognized for five minutes by seniority, and I will hold that
to five minutes today.
We have five votes that are scheduled for 11 a.m. I'll make
sure that all members today are able to ask these witnesses
their questions. I would ask members that, when you need to
leave to vote, would you be expeditious in coming back if
you're in the queue to ask questions, and the Chair will work
with each one of you to let you know where we think you'll be
in the sequence.
The Chair will announce he's going to miss the first two
votes to stay here and keep the continuity of the hearing going
so that we can get through as many members as we possibly can.
With that, Dr. Helmus, I'll recognize you and we'll go from
your right to left from there on. Dr. Helmus, the floor is
yours.
STATEMENT OF TODD HELMUS, Ph.D., SENIOR BEHAVIORAL SCIENTIST,
RAND CORPORATION
Dr. Helmus. Thank you, Chairman.
Good morning, Chairman Burr, Vice Chairman Warner and
distinguished members of the Committee. Thank you for the
invitation to testify at this important hearing.
Russia is engaged in a worldwide propaganda campaign. One
particular focus for this campaign is in Russia's own backyard,
in the former Soviet states of Eastern Europe. In addition to a
military and propaganda war in Ukraine, Russia is disseminating
propaganda to Russian speakers in the Baltics and other nearby
states.
Their goal principally is to drive a wedge between these
Russian speakers and their host nations, the North Atlantic
Treaty Organization, and the European Union. To do this, Russia
uses--Russia, of course, uses bot and troll social media
accounts. They also synchronize such tools with their state-
funded television network, their online news portals, and an
army of regional proxies that some call ``useful idiots.''
The RAND study I will talk to you about today sought to
better understand the nature and effectiveness of Russian--of
pro-Russia outreach on social media. By focusing on the region
that includes Estonia, Lithuania, Latvia, Ukraine, Moldova,
Belarus, our research team sought to help advance NATO's
defense of the Baltic states and shed light on how to combat
this issue around the globe. My written testimony highlights
the analytic methods and key findings from this--from our
report, but for today's testimony, I'll focus on our five key
recommendations.
First, there's a need to further develop analytic methods
to track and target Russian propaganda efforts. To take any
action against Russian social media operations, it is critical
to identify Russian bot and troll accounts and track their
activity in real time. This will require continued analytic
advancements so that computers can distinguish between
authentic social media chatter and the adversarial information
campaigns that are to come.
Second, it is important to highlight and tag Russian
propaganda. The approach by international organizations
involves frequently websites or e-mail alerts which reach only
fellow activists or members of the policy community. Instead,
the research team argues that it is important to highlight
Russian propaganda in ways that are much faster and target at-
risk audiences.
One example is Google ads could potentially help improve
the speed and targeting of counter-messaging. The approach uses
videos and other content embedded in Google search results to
educate people who search for Russian-born fake news on Google.
Third, expand and improve access to local and original
content. One challenge, particularly in the Baltics, is that
Moscow-controlled media, especially TV, is a dominant source of
information for many Russian speakers in the region. Policies
should not so much counter the Russian narrative as to displace
it with more entertaining and accurate content. The team argues
for training Russian language journalists, increasing access to
Russian language television programming such as Current Time,
and highlighting the authentic voice of local influencers.
Fourth, the U.S., NATO and the EU must do a better job of
telling their story. They should, for example, offer a
compelling argument for Russian-speaking populations to align
with the West or individual nation-states to which they belong.
NATO should also more effectively communicate the purpose and
intent of its infantry battalions now stationed in the Baltics.
Finally, there is a need to build resilience in target
populations. This will include long-term effort to implement
media literacy training and integrate such training into
classrooms. A public information campaign that can immediately
convey the concepts of media literacy and the risk of Russian
propaganda may also be necessary.
Thank you once again for inviting me, and I look forward to
taking your questions.
[The prepared statement of Dr. Helmus follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Chairman Burr. Thank you, Dr. Helmus.
Ms. DiResta.
STATEMENT OF RENEE DiRESTA, DIRECTOR OF RESEARCH, NEW KNOWLEDGE
Ms. DiResta. Thank you, Chairman, Vice Chairman and members
of the Committee, for giving me the opportunity to address this
body today. I'm Renee DiResta, Director of Research at New
Knowledge, and I study computation propaganda.
Disinformation, misinformation and social media hoaxes have
evolved from a nuisance into a high-stakes information war. Our
frameworks for dealing with them have not evolved. We discuss
counter-messaging, treating this as a problem of false stories
rather than as an attack on our information ecosystem.
We're in the midst of an arms race, in which responsibility
for the integrity of public discourse is largely in the hands
of private social platforms, and determined adversaries
continually find new ways to manipulate features and circumvent
security measures. Computational propaganda and disinformation
is not about arbitrating truth, nor is it a question of free
speech. It's information warfare, it's a cybersecurity issue,
and it must be addressed through collaboration between
governments responsible for the safety of their citizens and
private industry responsible for the integrity of their
platforms.
Malign narratives have existed for a very long time, but
today's influence operations are materially different because
the propaganda is shared by friends on popular social
platforms. It's efficiently amplified by algorithms, so
campaigns achieve unprecedented scale. Adversaries leverage the
entire information ecosystem to manufacture the appearance of
popular consensus. Content is created, tested and hosted on
platforms such as YouTube, Reddit and Pinterest; it's pushed to
Twitter and Facebook with their standing audiences in the
hundreds of millions, and it's targeted at the most receptive.
Trending algorithms are gamed to make content go viral.
This often has the added benefit of mainstream media coverage
on traditional channels, including television. And if an
operation is successful and the content gets wide distribution,
recommendation and search engines will continue to serve it up.
We're here because the Internet Research Agency employed
this playbook. Their operation began around 2013, continued
throughout the 2016 election, and even increased on some
platforms, such as Instagram and Twitter, in 2017. The
operation reached hundreds of millions of users across
Facebook, Twitter, Vine, YouTube, G+, Reddit, Tumblr, and
Medium. Websites were created to push content about everything
from social issues to concerns about war, the environment and
GMOs.
Twitter accounts masqueraded as local news stations,
WhiteHouse.gov petitions were co-opted, Facebook events were
promoted, and activists were contacted personally via Messenger
to take the operation to the streets. Twitter accounts and
Facebook accounts associated with the IRA remain active today.
The focus of the IRA campaign was to exploit social and
especially racial tension. Despite YouTube's claim that the
content found on its platform was not targeted to any
particular sector of the U.S. population, the majority was
related to issues of importance to the black community,
particularly officer-involved shootings. Hundreds of thousands
of Americans liked Facebook pages with names like Blacktivist,
Heart of Texas and Stop All Invaders.
The amount of explicitly political content that mentioned
the candidates in 2016 was small, but unified in its negativity
towards the candidacy of Secretary Clinton. In content that
targeted the left, this included messages aimed at depressing
the turnout, particularly among black voters, or painting
Secretary Clinton in a negative light compared to Jill Stein or
Bernie Sanders.
Only the social networks that hosted this campaign are
currently in a position to gauge its impact.
The IRA was not the only adversary to target American
citizens online. The co-opting of social networks reached
mainstream awareness in 2014, as ISIS established a virtual
caliphate across all social platforms.
The debate about what to do about that made it obvious that
no one was in charge. That confusion continues even as the
threat expands. The Wall Street Journal recently revealed that
a private intelligence company, Psy-Group, marketed their
ability to conduct similar types of influence operations to
impact the 2016 election.
Social platforms have begun to take steps to reduce the
spread of disinformation and deserve credit for doing that.
These steps, several of which were inspired by prior hearings
in this chamber, are a good start, but as platform, tactics and
protections change, determined adversaries will develop new
tactics.
We should anticipate an increase in the misuse of less
resourced social platforms. We should anticipate an increase in
the use of peer-to-peer encrypted messaging services. Future
campaigns will likely be compounded by the use of witting or
unwitting persons through whom state actors will filter their
propaganda. We anticipate the incorporation of new
technologies, such as video and audio produced by AI, to
supplement these operations, making it increasingly difficult
for people to trust what they see.
This problem is one of the defining threats of our
generation. Influence operations exploit divisions in our
society using vulnerabilities in our information ecosystem.
They take advantage of our commitment to freedom of speech and
the free flow of ideas. The social media platforms cannot and
should not be the sole defenders of democracy and public
discourse.
So, we recommend immediate action to identify and eliminate
maligned influence campaigns and to educate the public in
preparation for the 2018 elections. We recommend an updated
global IO doctrine, including a clear delegation of
responsibility within the U.S. government. We believe that
private tech platforms must be held accountable to ensure that
they're doing their utmost to mitigate the problem in our
privately owned public squares, and oversight is key.
Finally, we need structures and cooperation, information-
sharing between the public and private sectors. Formal
partnerships between security companies, researchers and the
government will be essential to defending our values, our
democracy and our society.
In closing, thank you for the opportunity to participate in
this conversation.
[The prepared statement of Ms. DiResta follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Chairman Burr. Thank you, Ms. DiResta.
Dr. Kelly.
STATEMENT OF JOHN W. KELLY, Ph.D., FOUNDER AND CEO, GRAPHIKA
Dr. Kelly. Chairman Burr, Vice Chairman Warner, members of
the Committee: Thank you for this opportunity to appear before
you today to discuss the weaponization of our social media
platforms and the resulting harm to our democracy.
The data now available make it clear that Russian efforts
are not directed against one election, one party, or even one
country. We are facing a sustained campaign of organized
manipulation, a coordinated attack on the trust we place in our
institutions and in our media, both social and traditional.
These attacks are sophisticated and complex, and the
Committee's bipartisan work to untangle and expose them sets a
great example for the country.
I am a social scientist and the CEO of a marketing
analytics firm that develops advanced techniques for
understanding the flow of information online. My experience
with Russian online communities began 10 years ago when I
helped lead a research effort at Harvard's Berkman Klein Center
for Internet & Society. In this work, we observed Russia's own
online political discussion evolve from a vigorously free and
open forum with a wide variety of organic voices and viewpoints
to a network rife with automated accounts and organized pro-
government trolling. In short, for the past several years the
Russian government has been doing to us what they first did at
home and in Eastern Europe a decade ago.
We know this because of indispensable work by a wide range
of investigative journalists, academic researchers, NGOs,
grassroots organizations, often conducted at great personal
risk. For more than a decade, these groups have documented the
playbook used by the Russian government to spread chaos and
discord online. These techniques include crafting fictitious
online personas to infiltrate communities, infiltrating radical
political communities on both sides to enhance their mutual
distrust, targeting both sides of a country's most divisive
issues, mixing pop culture references and radical political
discourse to influence young minds, using bots and trolls for
inorganic amplification, launching cyberattacks in conjunction
with information operations.
Again, each one of these features of the Russian
government's attack against the American public was first
tested and deployed against their own people, and then refined
to target their chosen enemies abroad.
Thanks to the great work of this Committee and to the
cooperation of social media platforms, data documenting the
Internet Research Agency's U.S.-focused effort in 2016 has now
been released to the public. Many dissertations will be written
on this data, but today I want to highlight just three points.
First, Russian manipulation did not stop in 2016. After
Election Day, the Russian government stepped on the gas.
Accounts operated by the IRA troll farm became more active
after the election, confirming again that the assault on our
democratic process is much bigger than the attack on a single
election.
Second, they are targeting both sides of our political
spectrum simultaneously, both before the 2016 election and
right now. We see from the IRA data how the same Russian
organization will use sophisticated false personas and
automated amplification on the left and the right in an attempt
to exploit an already divided political landscape.
Our current landscape is particularly vulnerable to these
sorts of attacks. In our estimate, today the automated accounts
at the far left and the far right extremes of the American
political spectrum produce as many as 25 to 30 times the number
of messages per day on average as genuine political accounts
across the mainstream. The extremes are screaming while the
majority whispers.
Third, American media is also being targeted. The IRA
persona ``Jenna Abrams,'' which had accounts on multiple
platforms, was cited by over 40 U.S. journalists before being
unmasked. The Russian activity seeks to turn the normal
differences of opinion among Americans into headlines about
unbridgeable political divisions. American journalism has a
responsibility to harden itself to these manipulations.
The platform's proactive transparency in these matters will
be critical in keeping us ahead of the new efforts and tactics
and to informing public debate on how to strengthen our
democracy in the face of these threats. There are significant
challenges ahead of us, and, unfortunately, knowing the other
team's playbook does not mean you are going to win the game.
The released data allow us to understand what the IRA did
in retrospect. Detecting these efforts before they have already
had their intended effect and agreeing on how to address them
remains a formidable challenge.
On the technological front, our field is making progress in
discerning technical markers that distinguish true grassroots
movements from fabricated campaigns. And research is yielding
methods for detecting manipulations before they gain momentum.
It is equally important to keep our values front and center in
this work, notably our dedication to freedom of expression and
to protecting user privacy.
It will take skilled women and men professionally dedicated
to this task and an investment in the development of tools and
methods to first catch up and then stay ahead in our race to
defend America's cyber social fabric from a new form of 21st-
century warfare.
Civil society or media institutions in the technology
sector can only do so much in the face of it. The
responsibility also lies with government to ensure that any
state actor eager to manipulate and harass faces consequences
for their actions. It is not just bots that are attacking us
and it's not just algorithms that must protect us.
The efforts of this Committee represent a tremendous step
forward in what will undoubtedly be a long and challenging
process, and I commend its leadership, dedication, thoroughness
and bipartisan spirit.
Thank you again for the opportunity to participate today.
[The prepared statement of Dr. Kelly follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Chairman Burr. Thank you, Dr. Kelly.
Ms. Rosenberger.
STATEMENT OF LAURA ROSENBERGER, DIRECTOR, ALLIANCE FOR SECURING
DEMOCRACY AT THE GERMAN MARSHALL FUND OF THE UNITED STATES
Ms. Rosenberger. Thank you, Chairman Burr, Vice Chairman
Warner and distinguished members of the Committee. I submitted
my full statement for the record, but let me highlight a few
key points on the national security context of these activities
and steps we need to take to address them.
The health and strength of our democracy depends on
Americans' ability to engage freely in political speech, to
hold vibrant debates free from manipulation, and to obtain
reliable information about the issues of the day.
I come at this issue as a national security professional
who has watched social media and online platforms be weaponized
to attack these foundations of our democracy. I watched from
inside the National Security Council when Russia test-drove
these approaches in Ukraine and as our government struggled to
understand them and respond. And I watched from the campaign
trail in 2016 as our government was surprised that these tools
were used against American democracy.
The 9/11 Commission characterized the failures that
preceded that attack as a failure of imagination. I believe the
failure to detect and disrupt the Russian government's
weaponization of online platforms to be a similar failure to
imagine, not just by the government but also by those who ought
to understand these tools best, their creators.
Thanks in part to the bipartisan work of this Committee, we
now know that Russian government-linked actors used a range of
means to manipulate the online information space, using nearly
every social media and online platform to amplify extreme
content and promote polarization, manipulate search results,
encourage action off-line, undermine faith in institutions,
insinuate themselves to target audiences in order to influence
public debates on geopolitics, and spread hacked information.
And, it's not just the Internet Research Agency. We know
Russian military intelligence officers used fake social media
personas and websites, and the United States is not the only
target.
The Chinese government has also begun to use social media
to manipulate conversation and public opinion outside of its
borders. Our authoritarian adversaries are using these
platforms because controlling the information space is a
powerful means to undermine democratic institutions and
alliances and advance their geopolitical goals. But meaningful
actions to close off these vulnerabilities by both government
and the private sector are lacking, and as we focus on the past
we are missing what still is happening and what will happen
again. What may have once been a failure to imagine is now a
failure to act.
Fundamentally, this is not a content problem. This is a
deliberate manipulation of the information space by actors with
malicious intent engaging in deceptive behavior. Transparency
and exposure of manipulation is critical to reducing its
effectiveness and deterring it, but tech companies have
remained defensive and reluctant to share information. Their
focus cannot be on public relations campaigns; it needs to be
on detailing the nefarious activities these companies are
seeing and curtailing it. Facebook's announcement yesterday is
what we need more of.
Transparency is also critical for accountability, and
outside researchers need greater access to data in a manner
that protects users' privacy. Users also need more context
about the origin of information and why they see it, including
disclosure of automated accounts while protecting anonymity.
Identifying malicious actors and their patterns of activity
requires new mechanisms for sharing data, both between the
public and private sectors and among technology companies.
Massive efforts along these lines are welcome, but need to be
streamlined and institutionalized and protect privacy and
speech.
We also need to identify threats in new technology before
they are exploited. AI presents new tools to both combat the
problem as well as new ways to make it worse, such as deep
fakes. Government and tech companies need to close off
vulnerabilities that are being exploited, including by
providing a legal framework such as the Honest Ads Act that
applies the same standards to political ads online that apply
off-line.
Manipulation of social media is one part of a larger
strategy to weaken our democracy. My bipartisan program
recently released a policy blueprint for countering
authoritarian interference in democracies endorsed by a
bipartisan and trans-Atlantic group of former national security
officials. Our recommendations include sending clear deterrent
warnings to foreign actors about the consequences for such
activity and identifying our own asymmetric advantages.
Government also needs to expose foreign interference
publicly, and legislating reporting requirements for the
Executive Branch would ensure that politics are not a
consideration.
We also need to harden our electoral infrastructures
through measures like the Secure Elections Act, as cyber
attacks remain a core part of Moscow's arsenal. More broadly,
the government needs a unified and integrated approach,
including through a counter-foreign-interference coordinator at
the National Security Council and a National Hybrid Threat
Center.
Finally, this is a transnational challenge and it is
essential that we work more closely with allies and partners to
share information about threats and collaborate on responses.
Distinguished members, there are steps that we can take
today to make our democracy more secure. We need to come
together across party lines and between the public and private
sector to address this challenge. Putin's strategy is to divide
Americans from one another in order to weaken us as a country.
In the face of this threat, standing together as Americans has
never been more important.
Thank you.
[The prepared statement of Ms. Rosenberger follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Chairman Burr. Ms. Rosenberger, thank you.
Dr. Howard.
STATEMENT OF PHILIP HOWARD, Ph.D., DIRECTOR, OXFORD INTERNET
INSTITUTE
Dr. Howard. Thank you, Chairman Burr and Vice Chairman
Warner, for the opportunity to testify on foreign influence
operations and their use of social media platforms.
My name is Phil Howard. I'm a professor at Oxford
University and Director of the Oxford Internet Institute, a
department at Oxford. My own area of expertise includes
political communication and international affairs. And at the
institute, I've been leading a project on computational
propaganda, currently funded by the European Research Council,
and something that--a research initiative that started with
support from the National Science Foundation in this country.
I began working on these questions in 2010, but the project
really grew in the summer of 2014, when the Malaysian Airlines
flight was shot down over Ukraine. And in Hungary, where I was
based at the moment, at that time, many of my Hungarian friends
got multiple ridiculous stories about what had happened. We
knew these came from Russian or Russian sources.
There was one story that democracy advocates had shot the
plane down because they thought Putin was traveling on
commercial from Amsterdam to Malaysia. There was another story
that Americans had shot the plane down because the U.S. had
stationed troops in Ukraine. And far and away my favorite was
the story of a lost tank from World War II that had come out of
the great forests of Ukraine that was confused and had shot the
plane down.
It was at that moment that we realized the thrust of
Russian propaganda was not so much about creating one counter-
narrative and placing that story amongst a public, but creating
multiple, sometimes equally ridiculous, stories and placing
those stories in a public. What we did not expect is that
Russia would turn this campaign strategy on America, on the
other great democracies in the West.
I'm going to say a little bit about what we've learned over
the last few years about the form of these computational
propaganda campaigns and give you a sense of what I expect for
2018 and perhaps the years ahead.
We coined the term ``computational propaganda'' because
this kind of disinformation is unique. It makes use of
automation; it makes use of the social media algorithms that
technology firms themselves have built. And it makes use of
those algorithms to distribute targeted propaganda. This
propaganda includes falsely packaged news, misinformation,
illegal data harvesting, hacking. There's a range of techniques
that goes into backing computational propaganda.
And there's three kinds of campaigns that tend to target
voters. There are campaigns to polarize voters on particular
issues. For example, known Russian social media accounts will
simultaneously promote political action by groups like the
United Muslims of America and the Army of Jesus, or encourage
African-American political activity around Black Lives Matter
and encourage others to support the Blue Lives Matter movement.
The goal is to get groups of voters to confront each other
angrily, not just over social media, but in the streets.
Second, there are campaigns to promote or discredit
particular senators, presidential candidates and other
political figures. Foreign-backed rumormongering is not new,
but it is strategically targeted in a way that is new.
Third and perhaps most worrying for democracy is that
campaigns, some of these campaigns, discourage voters from
voting. Voter suppression is a common messaging technique aimed
at voters whose support for a candidate a foreign government
might find unpalatable. For example, voters are often told that
voting day has been postponed, or that they can text message
their vote in, or that the polling station has moved when it
has not.
In the case of the United States, these campaigns are
ongoing. Months after the last major election in the U.S., our
team demonstrated that disinformation about national security
issues, including information from Russian sources, was being
targeted at U.S. military personnel, veterans and their
families.
During the President's State of the Union Address, we
demonstrated that junk news, some of which originates from
foreign governments, is particularly appetizing for the far
right, white supremacists, and President Trump's supporters,
though notably not small ``c'' conservatives.
Our team has completed recently a global inventory of the
number of governments managing these campaigns and, while many
of us talk about Russia, I would say that the original writ of
our research was to track what the Russians and Chinese are
doing in this domain. So far, we have not documented much
Chinese activity. We know they spend time working on voters in
Taiwan, they work on the Chinese diaspora. We believe they have
capacity, but as of yet they haven't set American voters in
their sights.
We have found in this most recent inventory that there are
48 countries in the world with large political parties or
government agencies running misinformation campaigns either on
their own voters or on voters in other countries. There are
seven authoritarian governments, aside from Russia, that spend
money in this domain.
And overall, I would say it's time for democracies to
develop their own cyber-security strategies. The time for
industry self-regulation has probably passed. And I'm grateful
for this opportunity to discuss the possibilities going
forward.
[The prepared statement of Dr. Howard follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Chairman Burr. Dr. Howard, thank you very much.
I am reminded, after listening to all of the testimony,
that the 1960s strategies of Russia were simple: If it's bad
for America, it must be good for us. And it seems like this is
rooted in the same foundational strategic vision that they had
then.
The Chair would recognize himself for five minutes. I'm
going to ask all of you to follow my chart over there. I just
want to get your comments relative to whether this is accurate
or not.
[The material referred to appears in the Supplemental
Material on page 163.]
Chairman Burr. The red line represents the Russian
activities of the IRA Twitter activity relative to outside the
United States. The blue line is U.S.-focused IRA Twitter
activities. What that shows is a huge spike up in the 2014-2015
timeframe, which was the invasion of the Ukraine.
The next two jogs of the lineup are between 2015 and 2016,
and that's the Crimea propaganda, and the regional politics in
Belarus specifically.
And then all of a sudden you see this spike in the blue
line in the United States. I think the fascinating thing here
is that the spike is in 2017 and 2018, which tells us--and
correct me if I'm wrong--the effort in 2017 and 2018 was much
more intense than the effort in 2015 and 2016 in the lead-up to
an election. Am I misreading that?
[No response.]
So, Dr. Kelly, let me ask you this: Is it possible for the
mainstream media today to run a story that was the creation of
an effort by the IRA, that had no factual basis, but over the
transition of how their strategies work, it gained enough
coverage, belief that people had read it, that it got so big
that it had to have been real? Is that possible?
Dr. Kelly. I believe it is possible. I think the goal of
these information operations over the long term is to condition
the public and to weave the network, so to speak, that later
you can use it to move any sort of story.
Remember, a key feature of propaganda--you know, if you're
running a propaganda outfit, most of what you publish is
factual, so that you're taken seriously, and then you can slip
in the wrong thing at exactly the right time. I believe that's
what they've done, is cultivate a set of sources as
authoritative with content that's often just about Kim
Kardashian. And then those people become credible, they become
cited in the mainstream media. And then at that point, they can
start to move anything they want through it.
Chairman Burr. And is it the individuals that contribute to
that theme that's on a social media platform, in many cases
Americans responding, that gives it credibility? And are they
knowing or unknowing as to what they're participating in?
Dr. Helmus, have you got a strategy on that?
Dr. Helmus. Certainly I agree that there's no borders on
social media. There's no borders on media today. So certainly
content that's disseminated by one source could easily get
picked up by another. It's our observation from looking at
Eastern Europe that there's fundamental issues with journalism
training and quality that can certainly lead to and exacerbate
that type of issue of, you know, bringing viral content that is
otherwise false or untrue into perceptions of reality.
Chairman Burr. Ms. DiResta, you said, and correct me if I'm
wrong, IRA pages stay active today.
Ms. DiResta. Yes, sir, I believe that's true; and Twitter
accounts that were associated with IRA botnets also appear to
be dormant today with the potential to be able to be turned
back on at some point.
Chairman Burr. So with all the efforts by the Justice
Department at targeting by the public acknowledgement and
indictment of individuals, the IRA has not gone away?
Ms. DiResta. No, sir.
Chairman Burr. Their capabilities--and comment on it if you
will--their capabilities relative to Facebook's latest
disclosure may have gotten significantly better.
Ms. DiResta. One thing that's a very big, significant
challenge is attribution. So we can attribute this to the IRA,
perhaps. I also read the same news that you read yesterday and
don't have any inside information there. My understanding is
they believe it was the IRA based on image similarities,
tactical similarities.
What they did change was they paid in, I believe, U.S.
dollars and Canadian dollars. So they are no longer paying in
rubles. They are probably no longer using IP addresses that are
tied to Russia; slight increases in operational security that
will make them more difficult to detect.
The other thing that is going to go along with that,
though, is as attribution is so difficult, particularly for
outsiders who don't have access to that kind of account level,
what we call metadata, is that other people will be able to run
the same playbook, perhaps making it look like an IRA operation
when it was conducted domestically.
Chairman Burr. Individual or a nation state?
Ms. DiResta. Individual or nation state, yes, sir.
Chairman Burr. Great, thank you.
Vice Chairman.
Vice Chairman Warner. Thank you all for your testimony. I
think a couple of things. One, we're still mostly just talking
about the IRA activity, as opposed to what we don't know in
terms of other Russian services' activities. And we do know the
IRA, with the revelations of yesterday, has gotten better.
And we're going to still need to figure out their
tradecraft. And one of the things we need from expertise like
you is I feel like even when the platform companies are moving
in the right direction, they're only doing it looking at their
own universe, their own platform, not the interrelationship.
I think, Mr. Kelly, you said something that was maybe the
single most stunning line of all the testimony, that in terms
of the political content, particularly on the extremes, that 25
to 30 times more of that content is being generated by bots and
automated accounts rather than individuals. Is that correct?
Dr. Kelly. Yes, Senator, that's correct. If you look at the
American political spectrum and, say, array a set of
politically oriented Twitter accounts along an axis where on
one side you've got those that only talk to people of their
own, you know, stripe, and on the other it's the other stripe,
and most Americans are in between, connected to some on the
right and the left, those on the either extreme of that network
are shouting with automated amplification.
Vice Chairman Warner. So with a lot of that automated.
Let me state for the record, we had some of this--I've had
conversations with you in the past. There are very appropriate
and effective roles for automated accounts and bots in certain
cases. But I guess what I would ask--I'll start with Ms.
Rosenberger and Dr. Kelly on this: Shouldn't we as human beings
have a right to know--maybe not make a judgment, but a right to
know whether the content that we're receiving is coming from a
human being versus an automated account; recognizing that there
is good value in some of the automated accounts?
Ms. Rosenberger. Yes, Senator. I believe that context about
information is absolutely critical for consumers of that
information to be able to evaluate it. When we talk about
critical thinking in media literacy, this takes on wholly new
characters when we talk about online content. And so having
information about the origin of information, about whether or
not that content is being served up through an automated
process, why users are seeing that kind of information, I
absolutely believe that's critical.
One thing I do think is important in this conversation is
that we ensure we protect the anonymity online, which is
essential for democratic activists in authoritarian states. But
I believe very deeply that there are ways to identify
automation without compromising the ability for users, real
users, to be anonymous.
Vice Chairman Warner. Dr. Kelly, do you want to?
Dr. Kelly. Well, we have to recognize that automation is
performing a lot more functions online that simply supporting
Russian propagandists. And the fact that it's doing so many
different things, some of which are, you know, call them green
things we like and some of which are red things we don't like,
makes it extremely hard, without being able to know who's
running that robot, to know who's using it for good or bad.
Vice Chairman Warner. Dr. Howard, did you want to weigh in
on this?
Dr. Howard. No.
Vice Chairman Warner. Could we analogize to the markets
where, with the huge advances around HFT and high frequency
traders--the markets, in terms of trying to make sure that
things didn't get totally away, put certain speeds bumps in
place. And if the market jolts one way or another, there are
these speed bumps that then allow in a sense human activity.
With the, again, 25 to 30 times automation, if there are
stories that are trending at an enormously rapid rate, that
might be trending because they've got this enormous amount of
automation driving that story, you know, could there be some
kind of time out so that you could, a company, or some entity,
could evaluate whether this is actual, not actual? Something
looks phony here, fishy here? Any of you on that comment?
Ms. DiResta. I think that the parallel to HFT is spot on. I
think that it's an issue of information integrity. And one of
the challenges that the platforms have had is believing that
they need to address the core of the narrative. And what we
should be looking for is addressing the dissemination patterns
that you're mentioning.
Vice Chairman Warner. I think that's really--go ahead, Mr.
Kelly.
Dr. Kelly. Well, one thing to keep in mind is that, again,
automation is running all kinds of things. So it's not just
pushing Russian propaganda. It's pushing legitimate American
political speech. It's also pushing pop music elements in, you
know, marketing around music. So automation is doing a lot of
things in different places.
Vice Chairman Warner. And I'll make the comment that it
doesn't come with good or bad attached. But I guess I just
think as a human being, I ought to have that knowledge of
whether that message is being promoted to me by a human being
or by automation.
And I know my time's up. I just want to come back, asking
Ms. Rosenberger on the next round of, you know, could we deal
with that protection of anonymity, but still put some geocoding
so that if somebody says Richard Burr from North Carolina, but
it's actually come from a different location?
Thank you, Mr. Chair.
Chairman Burr. Dr. Howard, did you have something you
wanted to add to that?
Dr. Howard. I just wanted to add that the other possibility
is to have these accounts self-identify with B-O-T, bot, in the
name. That kind of disclosure is what helps users separate the
good content from the bad.
Chairman Burr. Great.
Senator Risch.
Senator Risch. Well, thank all of you for coming here
today.
I think the takeaway from this, after listening to all of
this, is something that's troubled me from the beginning and
that is how difficult this is. We know the problem. We have bad
actors putting out bad information. The difficulty is how do
you segregate those people who are doing this from Americans
who have the right to do this?
I've looked at the stuff that--that, as everybody has, that
is part of this. But yet, if you took one of those pieces, any
one of them individually, and looked at it and said, we just
discovered who's doing this, it's John Doe in East Overshoe,
New Jersey, there's nothing illegal about it. It may be
disgusting. It may be untrue. It may be with a bad motive. But
there's nothing--indeed, it's protected by the First Amendment
of the Constitution.
So how do you separate that person from someone who is
doing the same thing, but coming from Russia, but whose motives
are to enhance Russia by pulling down America? How do you
police that?
And I think, probably, the question that Senator Warner
asked about putting a speed bump in so that somebody can
evaluate this. I mean, that kind of puts--I want to be the
evaluator, and I think most everybody does, and that's the
problem.
And then you talked about protecting anonymity. How do
you--how can you protect anonymity if you're going to actually
do something against someone who is doing something that we
don't want done?
These are extremely difficult questions. And I appreciate
all the kind things you've said about this is bipartisan, we
all need to come together, et cetera, et cetera. We all agree
with that, but how in the world do you do this? I mean, the
takeaway here has got to be that this is just an enormous, if
not an impossible, thing.
Mr. Helmus, your thoughts?
Dr. Helmus. Yes, I absolutely agree. I think that is the
fundamental question.
In our research, we identified upwards of 40,000 accounts
centered around Ukraine that are putting out vociferously anti-
Ukraine content. And ultimately, the crux is are these bad
actors that are doing this? Or is this a free--other actors
practicing what might otherwise be their free speech?
So, that's challenged our bot detectors. So, there are some
ways, and I'll defer to others on the Committee who can speak
to these, but there are bot detectors that are available that
can detect some types of content that mimic the characteristics
of bots. But it is an arms race. As developers develop ways to
detect bots based on either inhuman levels of content, the
timing of their tweets, or what have you, the producers of
those bots will then identify other ways of circumventing that
and staying covert. So, it's an arms race and I think it will
just require constant research and evaluation to develop and
update new techniques.
Senator Risch. Ms. DiResta.
Ms. DiResta. What you're describing is a significant
problem for researchers as well. And we look at information
operations, trying to gauge, again, attribution or whether this
is organic or not.
Senator Risch. But what do you do about it when you do get
the attribution?
Ms. DiResta. We try to look at the content. Has it appeared
elsewhere? Is it affiliated with past IRA operations? Or is it
coming from somewhere else? So, we look at the origin.
We look at the voice; the actors that are pushing the
content. Are they bots? Are they humans? Is there something off
about the bio related to past tweets? There's a number of
signatures there. And then, we look at the dissemination
pattern. Does it look like it's been artificially amplified? Is
it being run through accounts, or groups, or pages that seem a
little bit dubious?
We try to flag things for the social platforms as well. We
believe firmly in transparent communication, where we're
saying, this is what we're seeing, what are you seeing? They
have access to metadata and to account information and to e-
mail addresses, phone numbers, things that people have
registered their accounts with. That is also a significant part
of the investigation of the operation.
There is no easy answer to this question. This is the
primary challenge and this is where we see even influence
operations going towards laundering narratives, either through
the unwitting or through participants. That's a hard problem.
Senator Risch. The analysis that you're talking about is
you're looking for all of these things. But you'll find, I
assume, some actors that are, what we would consider, bad
actors, but yet, some actors that we would consider good
actors, whether it was a U.S. government operation or
something.
Who makes the determination as to who's a good actor and a
bad actor? That's what I really, really struggle with.
Ms. DiResta. And I think the----
Senator Risch. Dr. Kelly, why don't you get your two cents
worth in?
Dr. Kelly. Thank you, Senator.
So it's tractable to tell what's fake. It's harder, but
doable, to figure out who is behind it. And then you need to
understand who's behind it, tracking the landscape of threat
actors. That's where somebody is making a determination who's
against our interests and who doesn't matter. Then, once you
have that, you know, it's up to government and other
appropriate folks to figure out the response.
I think to do that detection in the first place requires an
enormous amount of data and sophisticated methods of analysis.
And it's not just data from one platform, so, it can't happen
only internally. It has to happen with data from multiple
sources, which then gets to your, I think, extremely important
questions about who makes these determinations and who has the
right to see that private data.
I think we have to look at a model that's like cyber-
security firms. So there are trusted industry partners that
everybody trusts, that they know are going to be secure in the
way they handle that data. We need some sort of a facility like
that where these advanced----
Senator Risch. Of course, this is different than cyber-
security, in that with cyber-security you don't want anybody
entering a private space, whereas with this you want everybody
entering. That to me differentiates the two.
My time is up. Thank you, Mr. Chairman. Thank you.
Chairman Burr. Senator Feinstein.
Senator Feinstein. Thanks, Mr. Chairman.
I want to thank Facebook for their move yesterday to delete
32 pages and 290,000 accounts on the basis that Russia and
other outside actors are continuing to weaponize social media
platforms. I'm very pleased that Facebook took this action, and
I hope that all social media platforms continue to actively
counter Russia's foreign influence campaign. I have no question
that it's going on, and I have no question that it is related
to more than just election interference.
Let me ask this question: Since the 2016 election ended,
how many IRA accounts have any of you found that are still
active?
Dr. Kelly.
Dr. Kelly. We've been doing some work on this. We went and
looked--I mean, that list of accounts is extremely valuable. We
looked for live accounts on other platforms using open source
research tools and we found a great deal of accounts directly
connected to the closed accounts, which were active across
numerous platforms.
Senator Feinstein. Can you put a number on it?
Dr. Kelly. Of the sample we've looked at so far, it's
roughly 28 percent of those accounts are connected to at least
one live account on a different platform. We also know that
those accounts were connected to numerous other Twitter
accounts and where--we think of this as what we have here is
the tentacle of an octopus, and we don't know how far out on
the arm of that octopus that tentacle has gotten.
Senator Feinstein. How about Russia's accounts?
Dr. Kelly. The Russian accounts evident in this data?
Senator Feinstein. Right.
Dr. Kelly. Well, presumably these are IRA accounts too and
presumably they have their own--you know, they've got a
tentacle wagging in Russia as well and I don't know how much of
their effort this represents.
Senator Feinstein. Does anybody else on the panel have a
comment on this subject matter?
Yes. Please, doctor?
Dr. Howard. Thank you, Senator. My comment would be that
it's the social media firms who have that information. We do
our best juggling probabilities and percentages to make best
guesses about what kinds of account. Some of these accounts
occasionally slip into Cyrillic and then slip back. There are
some giveaways. But it's actually the social media firms that
have the best data on this.
Senator Feinstein. Well, let me ask you this question.
Facebook has alleged that IRA activity on its platform alone
reached 126 million people and that doesn't include Instagram
or Twitter. What can you say about the extent to which the IRA
activity reached real Americans?
Dr. Howard. I can say that it was significant, yet also
concentrated in swing states.
Senator Feinstein. I'm sorry? Concentrated in?
Dr. Howard. Swing states----
Senator Feinstein. Swing states.
Dr. Howard [continuing]. During the 2016 election. So
particular states got more of this kind of content than other
states.
Senator Feinstein. And what was the time that you looked at
that to draw that conclusion?
Dr. Howard. It was from the beginning of the presidential
debates until through to a few days after Election Day.
Senator Feinstein. Have you looked at it now?
Dr. Howard. Not in the last few months, no.
Senator Feinstein. Can you estimate the number of Americans
touched by Russian-linked activity in this area?
Dr. Howard. No. That is very difficult to do.
Senator Feinstein. Can anybody?
Yes, please go ahead.
Ms. Rosenberger. No, I just wanted actually to add a small
data point to this, which is we spend a lot of time talking
about Facebook and Twitter but as Renee highlighted and others
have noted, this is a problem of the entire information
ecosystem. This is cross-platform. Reddit confirmed hundreds of
IRA-created accounts. Tumblr did it and in particular on
Tumblr, that platform was used to target the African-American
community particularly.
So, I think this is why it's so really difficult to
quantify in any meaningful way the reach of these activities,
because this is across the entire ecosystem, not to mention, as
others were highlighting, how this information gets picked up
and then transmitted and amplified through mainstream media
outlets.
Senator Feinstein. Let me ask you, when information becomes
a weapon, does anybody see any need to change the environment
to prevent this from happening?
Ms. DiResta. I believe that many of us were advocating
doing that when it became clear that ISIS had turned the
information ecosystem into a weapon. I believe that,
unfortunately, the dialogue between the government, the
platforms and researchers was not necessarily where it needed
to be. There were a handful of convenings that tried. There was
the Global Engagement Center that was established, that's now
tied up in some funding morass and we're not really clear what
the status of that is.
The tech platforms, about two years after the extent of the
ISIS operation became known, established the Global Internet
Forum to Counter Terrorism. To the best of my knowledge, that's
not staffed so much as it is a repository of hashed content so
that platforms can participate in takedowns.
To answer your earlier question with one other point, we
did see in the public House data set, when the House released
the ads, that the ads were both demographically and
geographically targeted. The number of people who saw that
content, only the platforms have access to that information,
but we could also gauge the number of followers that did follow
the Russia pages. And that was in the neighborhood of a couple
hundred thousand on the largest pages.
Senator Feinstein. Thanks. My time is up.
Thanks, Mr. Chairman. Thank you.
Chairman Burr. Thank you.
Senator Collins.
Senator Collins. Thank you, Mr. Chairman.
Dr. Kelly, you have a very profound statement in your
testimony. You said: Russian efforts are not directed against
one election, one party, or even one country. What are Russia's
ultimate goals? Is it to undermine the public's faith in
Western democracies and so weaken the bonds that unite us, that
there are opportunities for Russia?
Dr. Kelly. Yes, Senator, I believe that's exactly correct.
I think they have long-term strategic goals, which include
weakening Western institutions and faith in democracy and
traditional sources of information and authority. That's the
strategic goal. And then they have a lot of near, short-term
tactical goals, things like injecting hacked information to
sway a particular event or election, and they're doing that
activity all around their periphery and now here.
Senator Collins. Ms. DiResta, this is a question for both
you and Dr. Kelly. Both of you emphasized that Russian
manipulation did not stop in 2016. In fact, you, Dr. Kelly,
said that Russia stepped on the gas and increased its activity.
And Ms. DiResta, you said that Russian efforts increased
postelection to promote racial tensions in our country.
We imposed sanctions on Russia. They seem to have done no
good when it comes to this kind of activity. What can we do
beyond educating the public to counter Russia more effectively?
Ms. DiResta, I'll start with you.
Ms. DiResta. I would say that one of the things that we
need to do is to evaluate our information operations doctrine,
JP 313. I believe Senator Warner alluded to this in his recent
policy proposals. I think that addressing the scale and
sophistication of information operations is something that as a
government we've not really looked at that in quite some time
and perhaps that would be a good place for us to start.
Senator Collins. Thank you.
Dr. Kelly.
Dr. Kelly. I think there's a technical component, which is
to be able to effectively detect and attribute this activity so
you can authoritatively prove it's happening, and then you have
a more traditional toolkit of foreign policy measures to take
action.
Senator Collins. Dr. Howard, I want to get to something you
said, and that was you gave us several compelling examples from
your Hungarian experience where they received clearly false
stories that were intended to explain the downing of the
Malaysian airline. And what's interesting to me is, based on
Dr. Kelly's testimony, it isn't just the Hungarian press that
is being manipulated or infiltrated or controlled, but we've
seen evidence where America's media is also being targeted.
Dr. Kelly pointed out that the Russian persona of Jenna
Abrams, who had accounts on multiple platforms, was cited by
more than 40 U.S. journalists before being unmasked. How can
the media be more sensitive or more aware, more on guard to
being manipulated in this way?
Dr. Howard. Thank you, Senator. The United States actually
has the most professionalized media in the world. It's learned
certainly to evaluate their sources and no longer report tweets
as given. So I would say that in this country, the most
professional news outlets are already on the defense. They
already have ways to ensure that the quality of the news
product isn't shaped by these constant disinformation
campaigns.
I would say that the greater concern would be amongst the
media institutions in our democratic allies. I believe that the
Russians have moved from targeting us in particular to Brazil
and India, other enormous democracies that will be running
elections in the next few years. And while we still see
significant Russian activity, those countries have the media
institutions that need to learn, need to develop.
Senator Collins. Ms. Rosenberger.
Ms. Rosenberger. Thank you, Senator. I would just add that
this is not a problem that we've overcome. We have one example,
for instance, of an IRA-created Twitter account, the hash--
sorry, the handle was ``wokeluisa,'' that was tweeting in
particular to African-Americans, focused on the NFL take-a-knee
debate. There were IRA-created accounts tweeting on both sides
of that debate. But that Twitter account in particular, which
was active through earlier this year, appeared in more than two
dozen news stories from outlets such as BBC, USA Today, Time,
Wire, The Huffington Post, and BET.
So, this was about four months ago. So, we really do need
to make sure that this information is not getting laundered
into the broader ecosystem, which is part of the strategy here.
Senator Collins. And the issue there is when we read it in
a credible source, we're likely to believe it.
Ms. Rosenberger. That's exactly right. It gives it that
much more credibility.
Senator Collins. Thank you.
Chairman Burr. Senator Wyden.
Senator Wyden. Thank you very much, Mr. Chairman, and thank
all of you.
It seems to me for now and the foreseeable future,
protecting America's private data is going to be a national
security issue. Cambridge Analytica, like the Russians,
exploited Facebook's lax protections to abuse Americans'
information. I believe a significant part of the failure is the
fact that the Federal Trade Commission doesn't have the
authority or the resources to be a tougher cop on the beat. And
I'm going to be rolling out a plan to fix that in the weeks
ahead.
Now let me go to questions. Ms. DiResta, your testimony
referenced the Russian Facebook pages in 2016, targeting both
the right and the left. But you noted it was the pages
targeting the left that included not only content intended to
appeal to its audience, but also content intended to suppress
the vote and be critical of Secretary Clinton.
In your view, does the apparent Russian content released
yesterday by Facebook resemble the content the Russians used
last time to attract an audience on the left and among racial
minorities, which the Russians then used to suppress their
vote?
Ms. DiResta. Yes, sir, it does. There's a strong component
of cultural posts that appear in communities and pages
targeting minority voters: a lot of pride, pride-related
content, less news, more memes and that reflects what we saw
yesterday.
Senator Wyden. I appreciate that, because content targeting
I think is clearly going to be a big part of the challenge. The
public has got to be aware of it, because not all Russian
propaganda is going to get caught. And Americans are inevitably
going to read some of it, particularly if it's consistent with
what they already believe.
So I gather what you're saying, Ms. DiResta, is the public
has got to be alert to a repeat of the 2016 Russian playbook,
which was to attract an audience on the left, discourage them
from voting. And that could mean attacking Democratic
candidates, pushing the line, in effect so that the Russians
are trying to make it possible that our votes don't matter. Is
that essentially your concern?
Ms. DiResta. Yes, sir. There's a lot of efforts to push
intraparty divisions on the left.
Senator Wyden. Good.
Let me ask you now, if I could, maybe for you, Ms. DiResta,
Ms. Rosenberger, Dr. Kelly, about this concept known as down
ranking. My interest here is that for the social media
companies there's just a mismatch of incentives. The social
media companies, they want users and clicks and impressions,
and inflammatory and often false content creates that. So even
when the companies can't or haven't decided to identify a
certain account as either foreign or nefarious, they can still
downgrade the posts to limit their exposure. This is an equal
or worse problem with conspiracies and junk news as it is with
foreign influence.
So my question here would be for the three of you: Do you
think these down-ranking programs are effective? Are they the
kind of thing that ought to be considered part of the kind of
toolbox as we look to deal with this problem, Ms. DiResta and
the rest of you?
Ms. DiResta. Sure. So I think that there's sort of three
facets to the toolbox. There is remove, reduce, or inform.
Inform means to add additional context to a post. This is
Facebook's framework right now.
Reduce would be to do something like down-rank it, per the
question earlier about is it possible to inject just a little
bit of friction? This is where down-ranking could potentially
be used as a tool, as attribution and authenticity and
integrity are established, to reduce the reach of content.
And then remove is, of course, the more--the most extreme.
Senator Wyden. Would any of you like to add anything? Yes?
Ms. Rosenberger. I'd just like to note that we talk about
down-ranking, but we forget that up-ranking is also part of the
process. These platforms are not----
Senator Wyden. You're being way too logical.
Ms. Rosenberger. These platforms are not neutral pipes.
Senator Wyden. Right.
Ms. Rosenberger. Information is not being served up without
some kind of algorithm deciding, for most of the platforms,
without an algorithm basically deciding what is served up at
the top. So when we talk about down-ranking, we have to start
from the premise that up-ranking is baked into the cake. And so
then the question becomes: are these platforms actually somehow
prioritizing bad, malicious information, right? That, as we
know and as others mentioned in their testimony, gaming these
algorithms, whether that's on trying to get certain content to
trend or, frankly, getting certain content to rise to the top
of Google searches, something that we know that Sputnik and
RT----
Senator Wyden. I'm over my time. I just want to be clear,
as the author of Section 230, the days when these pipes are
considered neutral are over. Because the whole point of 230 was
to have a shield and a sword. And the sword hadn't been used
and these pipes are not neutral.
Thank you, Mr. Chairman.
Chairman Burr. Senator Blunt.
Senator Blunt. Thank you, Mr. Chairman.
So much of the activity we're looking at on the charts and
today is largely of the IRA. What percentage of Russian-linked
activity would you anticipate that the IRA represents? Is this
half of everything they try to do, 90 percent, 10 percent? Who
would have a sense of what we're not looking at when we're
looking at the IRA activity?
Dr. Kelly. We've looked at a number of different known
disinformation campaigns and we think these are--the IRA folks
are involved in a minority of them.
Senator Blunt. In a minority of them. Do you think that
would be the case here as well?
Dr. Kelly. I do. The only thing--the thing we don't know,
though, is how much of the IRA this is.
Senator Blunt. Ms. Rosenberger, do you want to comment on
that?
Ms. Rosenberger. I would just only add that we know from
Special Council Mueller's indictment actually of the GRU, there
is one section of that that notes that GRU operatives utilize
social media accounts and fake Web sites that they created in
order to spread hacked information and other kinds of
weaponized information.
So we certainly know that there are other actors. GRU is
probably better at hiding their tracks than the IRA is, and so
I think that just speaks to again how this is probably just one
tip of the iceberg of what we're looking at.
Senator Blunt. So, the early discussion clearly has moved
from what the Russians were paying for, which appears to be a
very small fraction of the impact they were having. Does
anybody disagree with that? That is clearly--and this IRA
activity may--is some fraction of the Russian activity in 2016,
2017 and into 2018. That would be--so I think the indictment,
the Mueller indictment, said that there were probably at least
80 IRA employees involved and millions of dollars involved in
that effort.
I don't know what--is that 5 millions of dollars or
hundreds, hundred million dollars? What kind of--what amount of
money do you think the Russians invested in this effort that
was covered by the Mueller indictment? He uses the term
``millions of dollars.'' That could mean a lot of different
things. Any idea of the activity you've looked at, what kind of
investment of money and how many people that may have been
involved in this?
Dr. Howard. We've done that audit globally. We believe that
half a billion dollars have been spent by the 40 governments
that we've studied since 2010. In the Russian case, we think
it's around $200 million U.S. over this extended period for the
full set of organizations behind the various campaigns.
Senator Blunt. Dr. Howard, on that topic, in the other
countries you've looked at, who should we be looking at after
Russia that are likely impacting our daily conversation in the
country, in some ranked order? Who would be the top three or
four countries that you would believe would be most actively
out there doing what Russia is also doing?
Dr. Howard. Well, in our research we look at Turkey, China,
Hungary and Iran.
Senator Blunt. Dr. Kelly, have a thought on that?
Dr. Kelly. We believe there's a growing black market for
people skilled in the--who have these dark arts, and they're
employing them in their own countries and they're also starting
to get hired to work in other countries. So, this is a critical
challenge, because the Russians may have been the first to
effectively do this, but they're not the only players; and
you'll have a black market of players who are mobile and can be
hired by any actor.
Senator Blunt. Well, just to be sure I understand, doctor,
the 40 countries, are these 40 countries you've looked at for
outside activity or 40 countries that are participating in this
kind of activity?
Dr. Howard. These are 40 countries that have organized
disinformation campaigns in the sense of stable personnel with
telephones and family benefits. These are formal organizations
that do this work.
Senator Blunt. And how many countries do you think they,
those 40 countries, would be trying to influence activity in?
Dr. Howard. Seven countries.
Senator Blunt. Seven countries?
Dr. Howard. There's seven authoritarian regimes that have
dedicated budgets for disinformation campaigns targeting voters
in other countries.
Senator Blunt. And how many other countries, again?
Dr. Howard. Our audit of government expenditures covers 40
in total. It's usually the United States, Canada, Australia,
the U.K. that are the--Germany--that are the targets.
Senator Blunt. That are the targets.
On Dr. Kelly's comment about determining the attribution,
you know, we have--in our country, we are focused on defense.
No administration has yet figured out what our offense should
be, and I think one of those reasons is we have not figured out
with certainty how we would determine where a cyber attack came
from as opposed to even cyber misinformation, which is a
different kind of cyber attack, but vulnerable infrastructure.
What we're seeing here is a vulnerable social media
infrastructure that may be every bit as critical infrastructure
as any of the other infrastructure we're trying to protect.
Ms. Rosenberger, I'm going to let you have the last answer
to my questions.
Ms. Rosenberger. Senator, I would just note on that, that
Russia is playing to its asymmetric advantage. This is a low
cost, high reward kind of tactic. We need to also evaluate:
what are our own asymmetric advantages and sometimes that's not
responding symmetrically or in the same domain.
So, for instance when it comes to Russia, I think this is
why imposing costs in the financial space in particular--we
know that Putin cares most about his power and his power rests
on his money. And I think that looking at ways that we can dry
up the sources of funding both for these activities as well as
for the regimes that are using them is incredibly important.
When it comes to China, things like reputational costs are very
important.
So I think, this is why it's important that we put this
conversation on the national security front in a broader
strategic frame to identify our own asymmetric advantages so we
can go on offense.
Senator Blunt. Thank you, Chairman.
Chairman Burr. Senator Heinrich.
Senator Heinrich. Ms. Rosenberger, I believe it was you who
said, and I may be paraphrasing here, but we've moved from a
failure of imagination, to a failure to act. Do you find it
troubling that, despite the current risk, despite the quickly
approaching 2018 midterms, that concrete responses like the
Secure Elections Act, like the Honest Ads Act, have not been
scheduled for a vote in the United States Senate?
Ms. Rosenberger. Yes, Senator. I do believe that, while
this is a complex problem, there are some clear steps that we
can take in particular on the defensive side, as well as on the
deterrent side, that we need to be taking urgently.
Senator Heinrich. I share that concern, because I think
some of these things are sitting right in front of us and we
just need to make it a priority.
For Ms. DiResta and Dr. Kelly: The Committee's analysis
shows that the Internet Research Agency's campaign focused
heavily on socially divisive issues, but fanning racial
division in particular was the single most targeted category of
effort. Are Russian information warfare operations using
unresolved racial tensions here as a weapon to weaken the
United States?
Ms. DiResta. Yes, I believe they are.
Dr. Kelly. Absolutely.
Senator Heinrich. Do you see that ongoing exploitation of
racial tensions as a direct threat to our national security
and, for that matter, our cohesiveness as a country?
Dr. Kelly. You could think of this as a social cohesion
attack to try and drive wedges into the American public where
maybe a little wedge or a piece of history in our past is being
exploited to make 21st century America look more like 1950s
America than it ought to.
Ms. DiResta. I would agree.
Senator Heinrich. So, we now know much more about the
Russians' 2016 campaign than we did before we started this
investigation, and we know it was far broader than we
originally thought. We know that it's highly active today, as
many of you have testified to, and we know that no single
entity by itself--not the government, the social media
companies, not civil society--can effectively stop foreign
influence operations on social media.
But, Ms. Rosenberger, in your view have we as a Nation
extracted the sort of price or penalty for this behavior that
would defer--deter Vladimir Putin from acting in this way? Or
has the Russian Federation simply gotten a pass so far in terms
of the price that we have chosen and that this Administration
has chosen to extract?
Ms. Rosenberger. So, I think it's evident by the fact that
this kind of activity continues, that we have not yet
effectively deterred it. One thing I would note is that in
classic deterrence theory, deterrence relies on two prongs: one
is credibility and one is capability.
And I think it's incredibly important, number one, that on
the credibility front we have very clear, consistent messages
from across the government, starting with our leadership and
all the way down, that they're----
Senator Heinrich. Including the White House?
Ms. Rosenberger. Including the White House--that this
behavior will not be tolerated and that there will be
consequences for it going forward, and articulating what those
consequences will be. And I think that there is a role for
Congress to play here in terms of teeing up triggers that would
be automatic, and I know there is consideration of such
measures and I welcome that. But I think that it also has to
start--the credibility piece has to be very, very clear.
Vladimir Putin cannot see from one place that there is a
potential for consequences, but then over here be getting a
very different mixed message. We have to have consistency; that
has to be credibility coupled with the capability to act.
Senator Heinrich. I could not agree more.
You mentioned financial cost as one of our asymmetric
advantages. What would you foresee as a potential cost that we
might extract for this kind of ongoing misbehavior?
Ms. Rosenberger. I think there's two different ways of
looking at it. One is, of course, very targeted sanctions and
other kinds of designations; the other is thinking more broadly
about how our financial system, the Western financial system,
frankly, is used for Putin and his cronies to hide the money
that they have stolen, by the way, from the Russian people.
And just as we have vulnerabilities in our information
domain, we have vulnerabilities in our financial system. I
think steps like providing transparency around beneficial
ownership, extending and legislating the geographic targeting
orders that the Treasury Department has been using--there's a
whole suite of steps that we outline in our report that I
mentioned earlier, that I think----
Senator Heinrich. I will read those in the report. I want
to hit one last thing and then my time is up.
You all mentioned the broader ecosystem. Can you just
confirm so that people understand, this isn't just a couple of
platforms? This is music apps, this is video games, this is
meme sharing. It's much broader than Twitter and Google.
Dr. Kelly. I would expect that they have people whose job
it is to figure out how to exploit every small new platform
that comes along.
Senator Heinrich. Thank you all.
Chairman Burr. Senator King.
Senator King. Thank you, Mr. Chairman. And I want to
thank--thank you, Mr. Chairman, for calling this, what I think
is a very important hearing.
And thank you all for all the information that you've
shared. I've been listening and came up with a couple of
conclusions. Tell me if I'm right. One is: there is a massive,
sophisticated, persistent campaign on multiple fronts to
misinform, divide and ultimately manipulate the American
people. Is that accurate?
Dr. Kelly. Yes.
Senator King. I wanted to hear ``yes'' because nods don't
go in the record.
[Laughter.]
Dr. Howard. Yes.
Ms. Rosenberger. Yes.
Senator King. Let the record show everybody nodded.
Dr. Howard. Yes, Senator.
Ms. Rosenberger. Yes.
Senator King. I think that's incredibly important because
in all of this whole Russia active measures thing, a lot of the
space and energy has been going into campaigns and elections
and collusion and those kinds of questions. This is an enormous
part of what's going on, and it worries me that we've sort of
lost sight of this.
The second thing I've learned from you is, number one, it's
still happening; is that correct?
Dr. Kelly. Yes.
Ms. DiResta. Yes.
Senator King. Absolutely, still happening?
Ms. Rosenberger. Yes.
Senator King. It's way beyond elections.
Ms. Rosenberger. Yes.
Dr. Howard. Yes.
Senator King. Secondly, it's more sophisticated than it was
in 2016. They're learning to hide their tracks, not paid in
rubles. I would have thought they would have figured that out
before. But more sophisticated.
And then finally, it seems to me what you've been
suggesting is we're asymmetrically vulnerable because of the
First Amendment and democracy. We believe--our whole system is
based on information. And we have this principle of opening
access to information. Thomas Jefferson said, ``We can tolerate
error as long as truth is free to combat it.'' Thomas Jefferson
never met Facebook, I might add.
But would you agree that we are particularly vulnerable
because of the nature of our society?
Ms. Rosenberger. Yes.
Senator King. Now, this one is for the record because I
think it's a long answer. It seems to me there are three ways
to combat this. And the first--and this is what I would hope
you would supply for the record--technical solutions. Things
that have been mentioned today that we could do, and that
Facebook could do, or Google, or Reddit, or Twitter, whoever.
Technical solutions: identifying bots, for example, those kind
of things.
Please give us some specificity and things that you think
we might be able to do without violating the First Amendment. I
shudder when I hear the words ``regulate the internet.'' I
don't want to do that, but there may be things that we can do
that could be helpful.
The second thing, it seems to me--and, Doctor Helmus, you
mentioned this in your testimony--we need to do a better job of
media literacy. I had a meeting just before, in the fall of
2016, with a group of people from Latvia, Lithuania and
Estonia. And I said, ``What do you do about this problem with
the Russians' propaganda? And you can't unplug the internet,
you can't unplug your TV.''
They had a very interesting answer. They said: ``The way it
works over here is, everybody knows it's happening and
therefore when something like this comes online, people say,
`oh, it's just the Russians again.' '' We haven't gotten to
that point.
Doctor Helmus, is that what you mean by ``improve media
literacy''?
Dr. Helmus. Yes, precisely. To be able to recognize these
instances when they appear, and to be able to process those in
a way that can minimize the impact.
Senator King. But that goes--it's deeper than just having a
hearing. This has got to be--you know, our kids are growing up
with these devices, but not necessarily being taught how they
can be manipulated by their devices. I think there ought to be
standardized courses in high school called ``digital
literacy,'' and increasing the public's awareness that they are
being conned, or that at least they're potentially being
conned, and how to ask those kinds of questions.
Ms. Rosenberger.
Ms. Rosenberger. Senator, I think that that's right; it has
to include online literacy as well as just your standard media
literacy. But it also can't just be in the schools. One of the
things we know from research is that, in fact, it may be that
older populations who are not growing up with technology may,
in some cases, be more vulnerable to manipulation by this kind
of activity.
Senator King. I would argue that's because they grew up
with newspapers and they have this unspoken assumption about
editors and fact checkers.
Ms. Rosenberger. I think that's probably right, sir.
Senator King. And if you do your website in Times New
Roman, people will give it some credibility.
Ms. Rosenberger. Especially if it's your friend sharing it,
or somebody you believe to be your friend, someone----
Senator King. And your friend may be sharing something
which they got from somebody that they didn't know where it
came from.
Ms. Rosenberger. Absolutely, absolutely.
Senator King. A final point, and I think you've touched on
this, is deterrence. Ultimately, we cannot rely exclusively on
defense. The problem thus far, it seems to me, is that the
Russians in this case and others see us as a cheap date. We are
an easy target with no results. Nothing happens.
And I would--that would be something I hope you all again
could take for the record because of a lack of time, to give us
some thoughts about deterrence. And I think it's important. It
doesn't have to be cyber. It could be deterrence in a number of
areas, including sanctions, as we've discussed.
But it has to be--there has to be some price to be paid.
Otherwise, as we now know, it's going to continue.
So give me some thoughts on deterrence for the record. I
appreciate it.
Thank you, Mr. Chairman.
Chairman Burr. Thank you, Senator King.
Senator Manchin.
Senator Manchin. Thank you, Mr. Chairman.
I want to thank all of you for coming today to help us.
This is a critical topic, which I hope all Americans are
watching.
We, as an open voting society need to be informed. A
properly informed voter population is the key to a sound
democracy. Unfortunately, Russia is trying to undermine that
foundation.
A quick look back through American history shows that our
allies and adversaries have changed over time. The Soviet
Union, specifically Lenin and Stalin, openly criticized the
capitalist West before World War II. During our mutual fight
against Nazi Germany, President Roosevelt called Stalin ``Uncle
Joe'' and the U.S. and USSR fought a mutual enemy. After the
end of the war, we found ourselves in an adversarial
relationship, known as the Cold War that lasted decades.
We saw a brief thaw in relations during the 1990s. But now
Russia, specifically Vladimir Putin, and the U.S. seem to be
adversaries again.
I would ask, I think Mr. Howard, in your written testimony
you describe Russian computational propaganda aimed at
everything that we've heard today, pulverizing voters,
discrediting certain political candidates, discouraging
citizens to vote.
So I would ask, which country--we know Russia--poses the
greatest threat to our democracy using social media platforms?
And which countries are making strides to do the same?
Dr. Howard. Thank you, Senator. I agree that Russia has
been the most innovative in developing these kinds of
techniques. Unfortunately, I think it's safe to say that
dictators learn from each other. So as they see successful
campaigns run in particular countries, they emulate. They sink
their own resources into developing similar capacity. Some of
these countries have re-tasked small military units to do
entirely social media campaigning.
So as I mentioned earlier, there are now seven different
countries that are--who are, most would agree----
Senator Manchin. Actively involved?
Dr. Howard [continuing]. Authoritarian regimes that are
actively developing these kinds of----
Senator Manchin. Which ones do you think--which one has the
greatest potential to do harm? Russia is unquestionably the
absolute greatest violator.
Dr. Howard. I believe China has the next best capacity in
this----
Senator Manchin. If they want to turn loose on us?
Dr. Howard. If they want to.
Senator Manchin. And you haven't seen that yet?
Dr. Howard. Not directly in the U.S. sphere.
Senator Manchin. I would ask this to any of you all. Is
there any country that has been successful at deterring Russia
or any other attackers from other countries?
Dr. Kelly. Not that I'm aware of.
Ms. Rosenberger. It's hard to know the counterfactual of
what would have happened in different cases in some of these
instances. There is some evidence that in the German and French
elections, that deterrent messaging from the top, from the
leadership there about the consequences for this kind of
activity, may have reduced in some ways the kind of activity.
Senator Manchin. How about Macron's election in France? We
saw that he fought back. As soon as they saw the attacks being
made by Russia, they were actively involved.
Ms. Rosenberger. There are some interesting lessons that we
may be able to learn from----
Senator Manchin. Dr. Kelly. I'm so sorry----
Ms. Rosenberger. No, please, absolutely.
Senator Manchin. Our time is very limited.
Dr. Kelly. No, I answered too quickly before. I think the
Macron case is a perfect example of how being aware of it, that
kind of situation awareness, as well as quick and decisive
action to counter it in terms of public--you know, speech by
the leadership--had an effect.
Senator Manchin. And let me just ask--I've got one final
question here. I have a little bit of time here, but I wanted
to see your all's opinion. In West Virginia, you know, people
are having a hard time deciding where to get the facts. And
fake news seems to be the real news, depending on where they
get it from, social media and sometimes on networks, if you
will.
Can I ask each one of you all, where do you receive your
news that you believe is factual? Where do you go to? Where
could I help a West Virginian find some real news and not have
to rely on trying to decipher themselves was it fake or not? Is
it made up, real or not?
And I'll start Dr. Howard and go right down.
Dr. Howard. I go to PBS, BBC, and the Canadian Broadcasting
Company.
Ms. Rosenberger. I'm old-fashioned and I tend to still like
newspapers as my sort of major sources. I like having
publishers involved and editors who are able to fact-check
content.
Dr. Kelly. I'm a New Yorker, and I'll go with the Old Gray
Lady.
Ms. DiResta. New York Times, Washington Post, Wall Street
Journal.
Dr. Helmus. Major newspapers.
Senator Manchin. Not one of you mentioned social media. Not
one of you all mentioned what we're here talking about as where
you get you news or where you trust your news to come from. I
think that speaks volumes of what we're dealing with today.
I have no further questions after that. Thank you very
much.
Chairman Burr. Thank you, Senator Manchin.
And I just might add to his comment about what happened in
France. France also did some things that constitutionally we
can't do. So let's recognize the fact that they had a very loud
message and they had a very big stick that they used. And we
might not get the same results, though that doesn't change for
the loud voice.
Senator Rubio.
Senator Rubio. Thank you.
No one mentioned TMZ. There is some good stuff on TMZ.
[Laughter.]
And I'm on as often as I can get on there.
Anyway, so I want to talk about the terminology that we use
because I think it's one of the things that's really impeding
the way forward, and get your insight on all of this. The first
is, I've had people come up to me and say: Well, everybody
spies on everyone. But this is not really about espionage,
certainly not in the traditional sense. This is not--I mean
there may be elements that involve espionage, you know hacking
a computer, getting into a system network and stealing e-mails
and the like. But this is not really an espionage situation.
The other term that's always thrown around is collusion.
And there's ongoing efforts to answer all those questions. But
this sort of thing doesn't really involve, or doesn't really
require collusion. You don't need the cooperation of a
political candidate or party to be able to do any of this.
In fact, many of the ads that were pulled down yesterday
have nothing to do with a candidate or a party in the short
term. And it isn't even quite clear what the psychology behind
it is, other than to get us to fight against each other.
So if you can just put--if people would just put aside the
whole espionage focus and put aside, you know, the collusion
focus, and let that be dealt with the way it's being dealt
with, we'd get left with the term ``interference.'' And that's
become such a generic term that it's almost become benign. You
know, ``interference'' sounds like everything from the
leadership of another country had a preference about who won
the election, to actually like actively engaged in helping
somebody get elected. And I would hope--and, maybe you
disagree--I hope you agree, this is more than that.
This is really, no, nothing less than informational
warfare. This is just another type of warfare to weaken an
adversary. And that's how Vladimir Putin views the United
States of America. So, for example, if he conducted a kinetic
strike, a military strike to take out anti-air defenses, he
would do so to weaken our air defenses. And if they conducted a
cyber attack to knock out our command and control, he's there
to weaken our communication systems or our electrical grid.
And if you do this, you do it in order to weaken our
society, our willingness and capacity to fight, to work
together, to come together as a Nation. This is part of their
broader doctrine on how to confront an adversary.
And on the escalation scale, it costs very little money,
you can do it with limited attribution, and it works because
the fact of the matter is, with all of the things happening in
the world today, the United States Senate Select Committee on
Intelligence has spent an inordinate amount of time on this
important topic and there are so many other issues we could be
focused on. So, it's worked to some extent.
Is this assessment of it right? Isn't this--this is not
interference. This is information warfare designed to sow
division and conflict and doubts about--because whether it
involves changing voter registration databases in the future at
some point, potentially, or the stuff we're seeing now, all of
that is designed to sow chaos, instability, and, basically, to
get us to fight against each other.
We're already fighting against each other in this country.
All this does is just, sort of, stir that up even more. Is that
an accurate assessment? Is this informational warfare?
Ms. Rosenberger. Yes.
Dr. Kelly. I agree 100 percent.
Ms. DiResta. Yes.
Dr. Howard. Yes.
Senator Rubio. So to the extent that it is--and I think
everybody's already asked you this question--but wouldn't one
of the best things that could happen is that--we can focus all
day on Facebook and Twitter, and Instagram. These are
ultimately platforms who are being used for informational
warfare. I don't believe they invited them in and there are
things they can do to improve their processes, and I wish their
disclosures were a little faster, but by and large, they're a
platform that's being used. It would be like blaming the road
builders because some enemy used that road that they built to
put their tanks into your country.
So there are things these folks can be doing to improve the
way they operate, no doubt about it. But ultimately, we really
should be focused on what's being done and not only who they're
using to do it.
And so my question is, why wouldn't these social media
pages be in a position to potentially alert all of their users?
Not just a public disclosure like they did yesterday in their
press conference but actively send out to all of its user's
alerts about every time they remove something, so that people
can become conditioned to the sort of messages that are being
driven by these informational warfare operations?
Ms. DiResta. I believe they can. I believe Senator
Blumenthal requested that they do so in response to the--back
in September after the first set of hearings. They did push
notifications to people saying that they had seen content, they
had liked a page, they had engaged. I believe Twitter sent out
e-mails to users who were affected.
That kind of disclosure is absolutely necessary, because
one thing that it does is it comes from a platform that is at
least seen as somewhat trustworthy, whereas if they hear it
from the media you see these polarized echo chambers where some
people don't even believe this is happening.
Ms. Rosenberger. Senator, I would just add that one of the
things we know from looking at both the history of active
measures as well as their use across Eastern Europe and Central
Europe is that sunlight is one of the most effective antidotes.
Transparency, exposure of this activity, is critical for both
building resiliency and deterring it going forward. And so, I
absolutely concur that the more information and the more
transparency that the platforms can be providing to their
consumers, to the users of information about these activities
is absolutely critical.
Senator Rubio. I don't have a question, Mr. Chairman. I
just want to say that it's great that Facebook put this stuff
out there and that we're having this hearing. I promise you,
the vast majority of people that I know back home will never
see a single one of these images because there's a lot going on
in the news every day, constantly, by the hour.
Chairman Burr. Senator Harris.
Senator Harris. Thank you.
Mr. Chairman, I'd like to put what I believe is a context
in which we should be thinking about what happened in 2016.
First, I think we're all clear that Russia attacked our country
during the 2016 election and that they are continuing to attack
us today. Russia not only attacked one of our most sacred
democratic values, which is a free and fair election, but also
I believe our very American identity.
I often say that we, as Americans, no matter our race,
religion, or region, have so much more in common than what
separates us. And among what we have in common is a love of
country and a belief that we as Americans should solely be
responsible for the choosing of our elected leaders and the
fate of our democracy and who will be the President of the
United States.
And I think of us then as being a large and diverse family,
the American family. And like any family, we have issues and
fissures that are legitimate and run deep and provoke potent
reactions. We have a history of slavery in this country. We
have a history of Jim Crow, of lynchings, of segregation, and
discrimination. And, indeed, we have a lot to do to repair and
to recover from the harm of the past and some harm that
continues today.
But let's be clear. Someone else came into our house, into
the house of this country, the family of who we are as
Americans, and they manipulated us; and they are an adversary,
and they provoked us and they tried to turn us against each
other. The Russian government came into the house of the
American family and manipulated us.
And we must take this seriously in that context and
understand that when we debate, as we did in 2016, one of the
most important debates that we have, which is who will be
leader of our country, the Russians exploited our Nation's
discourse to play into our deepest fear.
And as leaders I believe then it is incumbent on us to
speak to the American people about how we can solve this urgent
national security threat. I believe, first, we must act
urgently to bolster our country's defenses like our election
infrastructure and cybersecurity, a bipartisan issue that we
have been working on in a bipartisan way--I thank Senator
Lankford and many of our colleagues--throughout the work that
we've been doing on the Secure Elections Act.
But second, I believe we need to make sure that the
American public recognizes who is trying to sow hate and
division among us, so that the American public can rightly
identify and see it for what it is: an attempt to exploit our
vulnerabilities for the purpose of weakening our country and
our democracy.
And with that, I'd like to ask, Ms. DiResta, in your
written testimony you say that the Russian Internet Research
Agency, IRA, efforts targeting the right-leaning, quote,
``right-leaning and left-leaning Americans was unified in its
negativity towards the candidacy of Secretary Clinton''; and
that, quote, ``in pages targeting the left, this included
content intended to depress voter turnout among black voters.''
This seems to corroborate the intelligence community's
finding that Russia was trying to hurt the campaign of one
candidate in the 2016 United States election and help the
other. Can you tell us more about what your research has found
regarding the nature of the political content that the Russian
IRA was pushing toward Americans on social media during the
2016 campaign?
Ms. DiResta. It was unified on both sides in negativity
toward Secretary Clinton. It was not unified in being pro-
President Trump. So the pages targeting the left were still
anti-candidate at the time Trump.
On the right, we did see an evolution in which evidence of
support for candidate Trump continued during the primaries.
There was some anti-Senator Rubio, anti-Senator Cruz content
that appeared. And there was a substantial amount of anti-
Secretary Clinton content on both the right and the left.
On the left, that included narratives that either African
Americans should not vote, should vote for Jill Stein, which
was not a wasted vote, and during the primary there was support
for candidate Sanders.
Senator Harris. And then quickly, Ms. Rosenberger, you
recently published a report policy blueprint for countering
authoritarian interference in democracies. You described an
event on May 21 of 2016 where two groups were protesting in
Houston, Texas, and one was called the Heart of Texas that
opposed the purported Islamification of Texas. On the other
side, the United Muslims of America, who were rallying to
purportedly save Islamic knowledge, and these protests were
confrontational.
Can you tell me, at the time were law enforcement or the
protesters aware of who had manufactured the conflict?
Ms. Rosenberger. No, our understanding is that they were
not. One thing we do know is that, fortunately, law enforcement
was present at the demonstrations and therefore was able to
keep them separate. But one of the things that we believe may
have been part of the intent of organizing simultaneous
rallies--same day, same place, opposite sides of the street--
was probably to attempt to provoke violence.
Senator Harris. And then just quickly, if we can follow up
in any writing with the Committee, but I'd be interested in
knowing what your recommendations are for how we can inform law
enforcement, because obviously this is a matter that is about
public safety and frankly also officer safety. As we know, many
of these disruptions end up resulting in violence and harm to
many individuals.
Ms. Rosenberger. Absolutely. I would just point very
quickly to the announcement from Facebook yesterday, which
actually seems like it may have been something intended to be
along similar lines with a protest attempting to gin up very
high emotions.
Chairman Burr. Senator Lankford.
Senator Lankford. Thank you, Mr. Chairman.
To all of you, in your research and the data that you're
putting together to be able to help us in this and be able to
expose some of the issues, thank you. You all have done a lot
of hours at a computer and running a lot of data to be able to
get to this point. And we appreciate that very much.
Ms. Rosenberger, I want to ask you about some of the
recommendations that your team has made and to follow up on one
of the questions that Senator Blunt had started. You made some
very specific recommendations that, when we discover
attribution, which is not easy to do, but when we discover it
and see it as a foreign actor, three main sets of responses you
seem to have recommended: sanctions; making sure there's a
reputational cost for the country that's doing it; and
considering offensive cyber operations. I want to take those in
reverse order.
What would you consider an offensive cyber operation that
would be effective in this means?
Ms. Rosenberger. Well, Senator, as you know, the use of
offensive cyber operations is itself a very complex problem.
Senator Lankford. Right.
Ms. Rosenberger. So I'm just going to kind of boil it down
to be specific within this context.
What I would say is, I think that there are instances in
which when we are able to--when the U.S. government is able to
identify-- for instance, the servers that are being used to
carry out these operations, based on a variety of potential
damage assessments, et cetera, I do think that there are
instances in which that might be an appropriate course of
action.
Again, as we know in offensive cyber, this can often lead
to a challenge of whack-a-mole. You set up a new server, et
cetera. It does impose a cost. Of course, one of the things
that we know that creates challenges is sometimes for these
transnational operations they may, for instance, be using a
server in the United States, or in the country--or in the
domain of one of our allies. So that introduces complications.
So it's not a super-simplistic answer. But I do think that
there are instances in which we should consider it.
Senator Lankford. So you also mentioned reputational costs.
I'm not sure there's anyone left on the planet that doesn't
understand that Russia does propaganda on their own people and
does offensive propaganda against everyone else.
What kind of reputational cost could you put on Russia,
trying to expose their activities?
Ms. Rosenberger. Senator, the reputational cost
recommendation is a little bit more specifically aimed at
China, where I think that, as others have alluded to, China has
the capabilities and we're seeing them test these things in
their neighborhood. China has a longer-term strategic interest
that's much more about generating affinity toward it and its
model. So I think that reputational costs would be more
effective with China.
I concur with you that, when it comes to Russia,
reputational costs are difficult, although I do believe that it
is important for the American people to hear clear and
consistent messages from our leadership that Russia and
Vladimir Putin are an adversary and a threat to our Nation.
Senator Lankford. It was one of the areas that I was
pleased with Facebook's announcement yesterday that this
Committee had talked to Facebook about multiple times. It's one
thing to be able to say that they are being used by an
adversary; it's another thing to actually show the images.
Ms. Rosenberger. Yes.
Senator Lankford. Yesterday Facebook was rapid to not only
say there's an outside entity, we're not saying it's Russia, it
looks like it is, but here's the images they're putting out,
here are the events they're putting out. And they put out a
tremendous amount of data yesterday. That's much improved from
where we were two years ago, where they were still saying,
``We're not sure if they used us or didn't use us.'' Now
they're being very forward-facing on that. That's helpful to be
able to get information around faster.
Traditional media multiplied that message by putting it out
as well. That helps us to be able to get the message out.
That's one of the things that we heard on this Committee
multiple times: European allies have faced from Russia those
attacks, that they've been able to get that and have that
pushback immediately. So that was helpful to be able to see it
yesterday.
I have one other question to relate to this as well. You
had mentioned a comment here in one of your recommendations on
making sure that there is transparency, passing legislation
that ensures Americans know the source of online political ads.
Much of what happened with this was not an ad. It was just a
profile that was set up that they did a tremendous amount to be
able to develop it.
How do you separate out being aware where an ad is coming
from and just a profile that's a free profile, that's developed
quite a following?
Ms. Rosenberger. I completely concur that the political
advertising piece of this effort was a small one. My own view,
coming from a national security perspective, is when we
identify a vulnerability we should close it off. And so even if
it was not the most significant avenue that was utilized, I
absolutely believe that applying the same standards to
political advertising online that apply offline is absolutely
essential. That being said, that will not solve the problem and
we can't be in any way convinced that it will.
And so that's why we also recommend a number of
transparency measures about providing greater context for
users, about the origin of information, about whether
automation is involved, about requiring some kind of
authenticity confirmation while protecting anonymity. I think
these are the kinds of steps that can help mitigate some of
these broader concerns that you're raising.
Senator Lankford. I look forward to that conversation. We
also need to have a conversation on is there a level of
cooperation needed between the internet service providers, cell
phone companies, and others that have a different level of
information about where that information is coming from, and
their cooperation with some of the providers of content.
Right now we're leaning mostly on providers of content to
say, help us with the data and help police yourself on it. But
there's another whole level of information coming from the ISPs
and from the cell phone companies and such, as well, of where
that data is actually originating from.
Ms. Rosenberger. Absolutely. And when you combine that with
information that the intelligence community can provide, I
think that that is how we begin to put together different
pieces of this puzzle to create better identification
processes.
Senator Lankford. I look forward to that.
Thank you, Mr. Chairman.
Chairman Burr. Senator Reed.
Senator Reed. Well, thank you very much, Mr. Chairman, and
thank you all for your excellent testimony.
We all here appreciate what Facebook did yesterday. I think
it was a very appropriate and timely response. But there was a
comment that you made, Dr. Howard, that I think is very
important and bears repeating, which is basically that
companies are beyond self-regulation. Could you elaborate on
that, and then I'll ask the panel if they concur?
Dr. Howard. I think much of what we've discussed today has
come from evidence that has been released very slowly over a
two-year period, often after prodding from you, multiple kinds
of Committee investigations and multiple governments. When I
say that I think the social media industry is past the point of
self-regulation, I mean mostly that the more public, open data
there is about public life, the faster we can catch these
moments of manipulation.
For the most part, we've been speaking about American
citizens and us as individuals and the impact on our--on our
democracy, but democracies have civil society groups, faith-
based charities, civic groups, prominent hospitals and
universities that are also under attack. And these are also
distinct to democracy and these are part of--these are the
organizations that I think can help defend us.
Senator Reed. But I think, again, we have and we have
gotten--and Chairman and the Vice Chairman have done a
remarkable job. We've gotten, as you say, slowly and surely
we've gotten a little bit more response. But I think the time
is running out, frankly, and I think we have to move
legislatively to set in motion a framework of disclosure.
Someone mentioned, you know, options to remove information,
reduce information, or inform the participant. I don't think
that will happen voluntarily. It's the prisoner's dilemma. I'm
sure they would all love to do it, but unless everyone does it
it's not cost-effective or it's not culturally consistent with
their corporation.
So, let me just go with Ms. Rosenberger and down the line
about this comment about do we have to move very quickly to set
up the framework, consistent with the First Amendment
obviously, that allows us to deal with this issue?
Ms. Rosenberger. Senator, one thing I would note is that,
while the United States has not taken any steps like this,
other countries or international institutions have. So the
European Union has been moving out, not just with GDPR but
other conversations about regulation of social media and online
platforms. China is using its market access as enormous
leverage over these companies in order to basically set the
terms of the debates.
By being absent from this conversation and not taking steps
to figure out some of these very thorny issues, but right now
what's happening is other countries, other governments, are
setting the rules for this space. And that is in many cases not
in the interest of the United States. I think some of the ideas
that Senator Warner put forward in his paper earlier this week
are absolutely worth very, very serious conversations and the
kind of things we need to be doing.
Senator Reed. I think one of the ironies, as you point out,
is that we could be disadvantaged because not only don't we get
to make the rules, but our companies, our international
companies, will follow the rules in China, follow the rules in
Europe, and not follow the rules here, leaving us much more
vulnerable.
Dr. Kelly.
Dr. Kelly. I believe that it's critical to have access to
data from all the platforms in order to detect this kind of
activity. And that is a sophisticated analytic capability that
needs to be created, and it's going to be a lot of time and
effort from a lot of smart people.
Where does that data then sit? Who is it that gets to look
at it? And I think that our concerns about privacy and the
First Amendment lead us to at least suggest we ought to think
about industry-oriented consortiums or things that allow a kind
of--without moving it too far from industry--let them at least
have the first crack at the detection piece.
Senator Reed. Well again, I think your instincts are very
consistent with the views of most Americans. But this now has
been several years, and we are still waiting for the kind of
robust response. Perhaps the Facebook example yesterday is a
good sort of sign that the industry is coming around, but----
Dr. Kelly. Yes, Senator. I think that the proactive
transparency we saw yesterday from Facebook shows real
leadership in the field. And I think we need more of that.
Senator Reed. We do, and my concern is that, again, there
are other incentives, disincentives, profit, culture, et
cetera, that could inhibit that.
My time is expired, but ma'am, please.
Ms. DiResta. I think the key is to have oversight. We spoke
about finance a little bit earlier, high frequency trading in
particular. There were two sets of regulators. There were self-
regulatory bodies that stepped in, there were the exchanges.
There're some parallels there, where the exchanges are able to
see what's happening and immediately, before the regulatory
process happens, step in and say: Not on our platform.
I think that that's actually an interesting model; this
combination of regulatory, self-regulatory, the exchanges
acting independently, and an oversight body looking to make
sure the entire ecosystem remains healthy.
Senator Reed. You're talking about the security exchanges?
Ms. DiResta. Correct.
Senator Reed. Yes.
Doctor, comment?
Dr. Helmus. I'll just say our research certainly shows the
importance of tagging this information so that audiences can
know the source of it. The appropriate legislative mechanism
for that I can't speak to.
Senator Reed. Thank you.
Thank you, Mr. Chairman.
Chairman Burr. Senator Cornyn.
Senator Cornyn. I can't help but recall the words of H.L.
Mencken, who said that for every complex problem, there's a
solution that is clear, simple, and wrong. And so I think we
need to be a little bit--demonstrate a little humility when we
begin to approach this from a public policy perspective, what
our response should be.
But I also want to ask you about my impression, which is,
it would be a mistake to think this is just about elections.
And one of the reasons I say that, I came across an article
recently entitled ``When A Stranger Decides To Destroy Your
Life,'' where somebody used a fabricated story about a woman
and posted it online on a website called ``She's A
Homewrecker,'' and basically ruined this woman's life, or at
least challenged it in a dramatic way.
And then I thought, well, this is a tool that could also be
used by somebody who wants to tank a stock price by disparaging
the reputation of a company and then perhaps sell it short and
reap a significant reward. Or, if you're a Chinese telecom that
wants to get rid of some of the competition, particularly when
it comes to developing 5G technology or some other cutting edge
technology, this is also a pretty useful tool, using this
information warfare.
So all of this leads me to wonder if by focusing solely on
the election, which is dramatic and of tremendous concern--and
I share the concerns of all of you and all the Committee--that
if we just focus on that and not the rest of the picture,
whether we are missing the right picture.
Ms. DiResta, do you have any observations?
Ms. DiResta. Yes sir. We look at--at New Knowledge, we do
look at misinformation and disinformation targeting corporates.
On the state actor front, we have seen evidence of campaigns
targeting agriculture and energy as two industries of interest
to foreign powers. On energy, we've seen anti-fracking
narratives, anti-fracking bots, by countries affiliated--
countries with strong oil interests. In agriculture, that's
taken the form of spreading fear about GMO's.
Senator Cornyn. Yes, Ms. Rosenberger.
Ms. Rosenberger. I'd also note that in the case of Russia,
we know that they use these operations to try to shape our
conversations and views on geopolitical issues, especially
those of interest to Russia.
So for instance, one IRA-sponsored post on the fake--the
inauthentic account ``Blacktivist'' asked, how would we feel if
another country bombed us for the poisoned water in Flint and
for police brutality? That was posted in the immediate
aftermath of the Trump Administration's strikes on Syria after
the chemical attack in March of 2017. So a clear instance of
that account actually criticizing an action by the Trump
administration, using emotional issues like the Flint water
crisis and police brutality as an avenue in, to try to shape
views on a geopolitical issue of interest to Russia.
Senator Cornyn. Dr. Kelly.
Dr. Kelly. So I completely agree that there's a commercial
dimension of this which is underreported, and there's a lot
more going on in the commercial space in terms of these attacks
than is reported. Renee discussed some of them.
We've seen others with our customers. And sometimes they're
tied, these political attacks and the attacks on corporations,
where corporations will be basically punished with falsely
amplified boycott campaigns and similar measures for doing
something which is politically not what Russia would like to
see.
Senator Cornyn. Dr. Helmus, the psychologist Jonathan Haidt
gave a speech I saw online recently called ``The Age of
Outrage'' at the Manhattan Institute, where he basically
describes a narrative where there's a lot of things conspiring
to manipulate us and invoke outrage for whatever is going on,
whether it's cable news, social media or the like.
What can regular Americans do to protect themselves against
those, whether they be state actors, whether they be
individuals, whether with malicious intent? What can they do to
protect themselves? There's one thing for the government to do
what we can do from a policy standpoint, but what can average
individuals, consumers of social media online, do to protect
themselves from being manipulated by fake information or
misinformation?
Dr. Helmus. You know, our work, our work in Eastern Europe,
as was mentioned earlier, suggests that people in those areas
are very well aware of Russia's intentions. Russia lurks very
closely to those nations, and people know what's going on.
I think obviously the way to apply that to the United
States is understanding the need to know the sources of your
information, be able to adjudicate and assess the truthfulness
of that information, the potential biases of that information,
and then try to make your own decisions on that. Ultimately,
it's about being a careful consumer of information.
Senator Cornyn. Thank you.
Chairman Burr. Thank you, Senator Cornyn.
The Chair's going to recognize himself for just a question,
and then I'm going to recognize the Vice Chairman. We'll see if
we've got any members that return after that second vote
starts. But it's my intention to try to wrap up as close to
noon as we can.
You know, I've heard a lot of phrases to describe what went
on just in the last few minutes--disinformation campaign,
misinformation campaign, societal chaos campaign. Dr. Howard, I
think you used one that struck me earlier--computational
propaganda. And my suggestion is that we not come up with a
single one, because we're dealing with a generational issue.
And I think somebody alluded to it earlier, that it's much
easier to take a generation that grew up with these devices and
accomplished some type of change than it is for somebody that
struggled, like me, to learn how to use the device and found
the most useful TV ad, when somebody defriended somebody they
took their picture off the wall, if you remember that Post-It
note? That struck home to me.
So I think it's important that we speak to as many
languages on this, because the task that we've gotten before us
is to penetrate the entire population. And it's not limited to
the United States. As you have described today--and, you know,
I hope if there's a takeaway for the media--this is going on
everywhere. It's not limited to politics. It's much more
intrusive in the economic, global economic picture today, than
it is in the political landscape.
It's just we like to write about politics. And so, I want
to point you to this chart I've got over here. It looks like
something that would be used at the psychiatrist's office, to
have you describe what it was. And I'm going to ask you, Dr.
Kelly. In our analysis, we went through and we tried to connect
the dots: Who generates it, where does it go, does it go to the
right, does it go to the left? And what my staff determined--
and I'm looking for your agreement or disagreement--is that in
a lot of cases, at least in the '16 cycle, the same person
sitting somewhere in the world generating, initiating this
propaganda, both initiated the part on the right and the part
on the left, that it wasn't two different individuals.
Therefore, this was a very well-orchestrated, very
choreographed plan that they carried out.
What's your comment on that?
Dr. Kelly. Well, this is very interesting and it tells a
deeper part of the story that the Clemson--recent Clemson paper
tells, which is that you don't just have, you know, one room
full of people who are running right-wing trolls and another
room full of people running left-wing trolls. It's actually the
same people at the same computers. So, I think that is a real
lesson in how we need to worry about the way they're trying to
play us like marionettes, right and left.
Chairman Burr. And is it safe to say that it's so easy that
Russia uses existing views inside of American society; all they
do is try to make the gap bigger between the two by inflaming
both sides?
Dr. Kelly. I agree. I think that they're not creating these
divisions. They're not--you know, and they're doing the same
thing in Europe and elsewhere. They find in a society what are
the vulnerabilities, what are the groups that oppose each
other, and they're basically arming them. It's kind of like
arming two sides in a civil war so you can kind of get them to
fight themselves before you go and have to worry about them.
Chairman Burr. So, Ms. Rosenberger, is this any different
than really what we faced in the 1960s in the campaigns by the
Soviet Union against their adversaries in the world of
propaganda?
Ms. Rosenberger. It is and it isn't. I think the playbook
in some way is the same, but the tools that they can use to run
those plays are very different. And what we have seen is that
digital platforms have supercharged the ability to take that
playbook and to really reach a much broader audience more
quickly and in a much more targeted kind of way than what we
would have seen in the 1960s.
There's a difference between hand-cranking out leaflets in
a basement and passing them around under covert means than
there is from putting information online using automated
techniques, inauthentic personas, to watch it go viral.
Chairman Burr. I will say that the Vice Chairman has been
one of the most outspoken about how technology allows this plan
to be on steroids. Words like bots, and he comes up with some
new ones every day, that many on the Committee and most in the
country either didn't understand at the beginning of this or
still don't understand.
So I'm not sure that we can emphasize enough the intent,
but, more importantly, the capability, and he deserves a
tremendous amount of credit for raising this to the level that
it is.
I recognize the Vice Chair.
Vice Chairman Warner. Thank you, Mr. Chairman. That's the
nicest thing you've said about me and you said it with no
members here.
[Laughter.]
Chairman Burr. I can repair the record.
Vice Chairman Warner. You can repair it.
Well, I want to start with what Senator Cornyn and you just
said. I think the political piece of this is really going to be
relatively small compared to the overall threat. And I think
one of the things we've not talked about yet today is the
marrying of cyber attacks with misinformation and
disinformation.
So, if somebody goes out, and let's say, for example that
the Equifax hack was actually done by a foreign actor, and it's
got personal information on 146 million Americans, then that
actor contacts you with your personal financial information,
you're going to open that, open that message. And then, if
behind that messages comes a live-stream video of what appears
to be Mark Zuckerberg or Jay Powell, the Chairman of the
Federal Reserve, the ability to wreak havoc in the markets, it
really almost overwhelms what we've seen on the political
front. So this cyber-misinformation combination is one that's
important.
I appreciate when we were talking earlier and recognize the
rest of you--you really helped me recently--that even something
that seems so obvious as should we have the right to know
whether we're being contacted by a human being or a bot has
layers of complexity to it. But I think we ought to continue to
explore that.
Ms. Rosenberger, I've got two points I want to make. One
is: you have rightfully said we want to make sure that we
protect anonymity, particularly, you know, the foreign
journalists in Egypt or the female journalists in Egypt, and
the ability to hide sourcing gets easier and easier with the
use of virtual private networks.
Even with those challenges, shouldn't we have some ability,
though, to say if--should an American have some ability to put
some kind of geocoding location so that if somebody says
they're posting a message from Michigan or North Carolina and
it's originating in Macedonia or Russia, you ought to at least
have that information? Again, you can still--we don't have to
get to content, but we can just know that there ought to be a
second look, because the origin of that post may not be what is
described in the post.
Is that a possible tool?
Ms. Rosenberger. I think that there are ways that can be--
that's one thing can be investigated. I think there are a
variety of ways to require authenticity without requiring
disclosure, sort of frontally, right? So a platform--in fact,
some of them actually do require confirmation of authenticity.
Some of them require--some of them include a verified check
that then sort of puts another label of--another level of sort
of authenticity on top of that.
But I think that there are ways that authenticity can be
confirmed or at least we can do a lot better to try to confirm
it, while still ensuring that we do have anonymity protected
and--sorry.
Vice Chairman Warner. Let me follow up on that, because
we've heard today some members talk about Section 230. We've
heard some members talk about GPDR and the whole privacy
bucket. You know, I've raised some issues about humans versus
bots. We're talking here about geocoding.
One of the areas that we haven't talked so much about--and
I'll appreciate the Chairman giving me this extra time--but are
there market forces that could help regulate if we ensured more
competition? For example, I was an old telecom guy and it used
to be really hard to move from one telco to another until we
implemented requirements of number portability.
You know, the Facebooks, the Googles, the Twitters dominate
the markets. There may be, as people increasingly have concerns
about the safety of their data, the ownership of their data,
fake accounts being used--and this doesn't completely work as
an analogy; let me state that up front. But the notion of data
portability, the notion that would say: if you want to take all
of your content off of Facebook, including your cat videos,
they have to make it in a user-friendly form to move to NewCo,
because NewCo as part of their business model is going to have
much higher levels of authentication.
I mean, is this--is that a possible avenue to look at, as
well? And I'll take anybody on the panel. Now, and when you get
into data portability, you've also got to get into
interoperability issues, which makes it again not a perfect
analogy. But is there a nub of an idea there? Anybody?
Dr. Kelly. I don't have an answer on that exactly, but I
think as you're thinking about that it's important to think
that, in these kinds of disinformation campaigns, two of the
most powerful things are a combination of anonymity and
atomization. You know, those two things together allow you to
run very large bot armies, so to speak, that are able to effect
your objectives. It's important--so those two pieces are
something you have to think about, how that concern weaves
through this.
The other thing to realize, though, about that is that the
bots are only part of the army, so to speak. So by solving that
problem, even if you force them to identify, you've basically
forced a medieval army to, you know, put a flashing light on
the archers. There's a lot of other folks out there that are
playing more direct roles that you still have to worry about.
And I think that those more high-value assets in this kind of
cyber social battle are a little bit harder to find. And
they're the ones that, you know, you can't just fire up
another--another hundred of them if you shut--if you shut the
first one----
Vice Chairman Warner. We've done a lot of recognition of
Facebook today. I think we should also recognize Twitter, which
in the last two months has, you know, even counter to their
business model, has taken down lots of fake accounts, lots of
fake bots.
But is there any, you know--is there any possibility here
about trying to add more competition into the marketplace as a
way to help us sort through this? Not so much just a regulatory
approach, but a competitive approach?
Ms. DiResta. I'd say one of the challenges is if you
fragment the platforms and fragment where people are, then
there are more platforms to watch, since this is a systems
problem and it does touch everything. That's not to say that
that's not an appropriate course of action, because one of the
reasons why this is so effective is there is this mass
consolidation of audiences as the internet, which was
originally much more decentralized, kind of came to have mass
standing audiences on a very small handful of platforms.
The challenge there is also, though, that people like that
consolidation. They like having a lot of--you know, all of
their friends on one platform. So this is a--it's kind of a
chicken-egg problem to think about it in those terms, but happy
to continue the conversation.
Vice Chairman Warner. I would just--if anybody wants to
add, my last comment would be: I think one of the earlier
statements that were made was that each of these platforms,
even as large as they are, really only look after their own
content or their own usages. So that ability to see across the
whole ecosystem is mostly lacking.
And I think the Chairman and I--and we spent a lot of time
trying to learn up on this--feel like the U.S. government is
trying to get a handle on this, but has got a lot of work to
do, as well.
So I really want to thank all of you. And one of the things
that we might be able to find consensus on, you know, is there
more ability for us to urge, force, nudge the platforms in an
anonymous way to share more data with independent researchers?
Because you guys actually can give us that system-wide view
that, for all the size Facebook has, Facebook can't give us the
complete picture.
Ms. Rosenberger. Senator, I think that that's exactly
right. I think we need two different kinds of information
sharing, and ideally, they can somehow be combined. One is
greater data-sharing between the public sector and the private
sector, bringing together the capabilities of the U.S.
government and the intelligence community, with the
capabilities and what the platforms are able to see happening
in their own ecosystem. Of course, that needs to be with
privacy and speech protected, but I think there are mechanisms
to do that, number one.
Number two is cross-platform information sharing. So I
would think about this as both a vertical and a horizontal
challenge. And then you have the question of outside
researchers, which is absolutely critical. I think that Renee
mentioned earlier the Global Internet Forum to Counter
Terrorism. I think that's one model to look at in this space.
There's other models, including from the financial
integrity world as well as from the cyber security world, where
you have been able to bring together different parts of
industry, academics and the government to ensure that the full
picture is put together to best go at this problem.
Vice Chairman Warner. Well I just want to again thank all
of you, but I also particularly want to thank the Chairman, his
notional idea. He did get this beyond taking the Post-It note
off the refrigerator. But he has been a great ally, has moved
this Committee forward on a whole host of technology issues.
This is one where there is no Democratic or Republican
answer, since clearly the goal of our adversaries was not to
favor one party over the other. It was to wreak havoc and split
divisions. And I think this Committee, under your leadership,
is trying to take this issue on in an appropriate way.
Thank you, Mr. Chairman.
Chairman Burr. And I thank the Vice Chairman.
You know, I was just sitting here thinking a lot of good
has happened since we started this drumbeat over a year ago. A
lot of changes have happened that I think 12 months ago at some
of the companies we would have said ``Never do.''
A big ship is not turned around overnight. It takes a
while. But I think that they have now given us an opportunity
to work with them. And I hope that in a month, when we have at
least three of the platforms in, that we will see a willingness
to collaborate with us, to come up with a solution that fits
both legislatively and from a standpoint of their corporate
responsibilities.
So I'm optimistic that we're headed--that we're started on
that pathway to a solution. You know, I remind people that it
was this Committee that took on legislation for cyber security
when everybody said it couldn't happen. Is it perfect? No. Was
it a good first step? Yes.
And part of the challenge, because we're the filter for
technologies changes in the world--there's no Committee of
Technology in Congress, there's no Agency of Technology in
Washington. It all sort of dumps in our lap, and we have a
perspective that nobody else has. And technology will drive,
for the next 10 years, the way we do things, the way we
communicate, where we go, how we do it. Everything in life is
going to be driven by technological change.
So, this is very appropriate that we would be talking about
a new architecture, not necessarily a new architecture for
social media, but a new architecture for the relationship
between government and the private sector.
And I hope that if there's a takeaway from today's hearing,
it's that this is the last time we're going to associate the
propaganda effort that we see, with an election cycle. There's
been no interruption since 2016. There was no interruption from
2014. This was planned out well before we knew who two
candidates were, we knew the differences between two parties,
or where the American people's hot button was. It's flexible
enough and it's nimble enough that it's going to attack
whatever the hot button is at a given time that they want to
initiate.
I can't thank all of you enough for your candid and
insightful testimony. You've given us a lot to think about as
we wrestle with how to counteract the problems of foreign
influence and its use on social media.
I want to summarize what we've heard today for the American
people. The Russians conducted a structured influence campaign
using U.S.-based social media platforms and others to target
the American people, using divisive issues such as race,
immigration and sexual orientation. That campaign is still
active today. They didn't do it because they have political
leanings to the right or to the left, but because they--or
because they care about our elections--but rather because a
weak America is good for Russia.
Some feel that we as a society are sitting in a burning
room, calmly drinking a cup of coffee, telling ourselves this
is fine. That's not fine, and that's not the case.
We should no longer be talking about if the Russians
attempted to interfere with American society. They've been
doing it since the days of the Soviet Union and they're still
doing it today. The pertinent question now is: what are we
going to do about it? And it won't be an easy answer. The
problem requires all of us--government, private sector, civil
society, the public--to come together and leverage our distinct
strengths and resources to develop a multi-pronged strategy to
counteract foreign attacks.
We've heard about the problem today and have considered
some potential recommendations and solutions. The next step is
to hear from the leaders of social media companies themselves.
And I'm certain that they, too, learned a fair amount today
while watching this hearing, and I look forward to their
responses. They owe it to the American people to communicate
clearly and transparently what they view their role to be, and
what they're doing to combat these foreign influence
operations.
As I mentioned previously, this issue goes far beyond
elections. We're fighting for the integrity of our society. And
we need to enlist every person we can.
With that, I want to thank you for your time today. I think
I've hit within about a minute of what I told you our target
would be. This hearing is adjourned.
[Whereupon, at 11:58 a.m., the hearing was adjourned.]
Supplemental Material
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
[all]