[Senate Hearing 118-581]
[From the U.S. Government Publishing Office]
S. Hrg. 118-581
THE NO FAKES ACT: PROTECTING AMERICANS
FROM UNAUTHORIZED DIGITAL REPLICAS
=======================================================================
HEARING
BEFORE THE
SUBCOMMITTEE ON
INTELLECTUAL PROPERTY
OF THE
COMMITTEE ON THE JUDICIARY
UNITED STATES SENATE
ONE HUNDRED EIGHTEENTH CONGRESS
SECOND SESSION
__________
APRIL 30, 2024
__________
Serial No. J-118-64
__________
Printed for the use of the Committee on the Judiciary
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
www.judiciary.senate.gov
www.govinfo.gov
__________
U.S. GOVERNMENT PUBLISHING OFFICE
59-363 WASHINGTON : 2025
-----------------------------------------------------------------------------------
C O N T E N T S
----------
OPENING STATEMENTS
Page
Coons, Hon. Christopher A........................................ 1
Tillis, Hon. Thom................................................ 3
Prepared statement........................................... 38
WITNESSES
Barnett, Tahliah Debrett (``FKA twigs'')......................... 7
Prepared statement........................................... 48
Responses to written questions............................... 97
Crabtree-Ireland, Duncan......................................... 8
Prepared statement........................................... 50
Responses to written questions............................... 101
Davies, Graham................................................... 12
Prepared statement........................................... 57
Responses to written questions............................... 105
Kyncl, Robert.................................................... 5
Prepared statement........................................... 65
Responses to written questions............................... 109
Ramsey, Lisa P................................................... 13
Prepared statement........................................... 68
Responses to written questions............................... 115
Sheffner, Ben.................................................... 10
Prepared statement........................................... 83
Responses to written questions............................... 131
APPENDIX
Items submitted for the record................................... 37
COMMITTEE ON THE JUDICIARY
RICHARD J. DURBIN, Illinois, Chair
SHELDON WHITEHOUSE, Rhode Island LINDSEY O. GRAHAM, South Carolina,
AMY KLOBUCHAR, Minnesota Ranking Member
CHRISTOPHER A. COONS, Delaware CHARLES E. GRASSLEY, Iowa
RICHARD BLUMENTHAL, Connecticut JOHN CORNYN, Texas
MAZIE K. HIRONO, Hawaii MICHAEL S. LEE, Utah
CORY A. BOOKER, New Jersey TED CRUZ, Texas
ALEX PADILLA, California JOSH HAWLEY, Missouri
JON OSSOFF, Georgia TOM COTTON, Arkansas
PETER WELCH, Vermont JOHN KENNEDY, Louisiana
LAPHONZA BUTLER, California THOM TILLIS, North Carolina
MARSHA BLACKBURN, Tennessee
Joseph Zogby, Majority Staff Director
Katherine Nikas, Minority Staff Director
Subcommittee on Intellectual Property
CHRISTOPHER A. COONS, Delaware, Chair
MAZIE K. HIRONO, Hawaii THOM TILLIS, North Carolina,
ALEX PADILLA, California Ranking Member
JON OSSOFF, Georgia JOHN CORNYN, Texas
PETER WELCH, Vermont TOM COTTON, Arkansas
MARSHA BLACKBURN, Tennessee
James Barton, Democratic Chief Counsel
Geoffrey MacLeay, Republican Chief Counsel
THE NO FAKES ACT:
PROTECTING AMERICANS FROM UNAUTHORIZED DIGITAL REPLICAS
----------
TUESDAY, APRIL 30, 2024
United States Senate,
Subcommittee on Intellectual Property,
Committee on the Judiciary,
Washington, DC.
The Subcommittee met, pursuant to notice, at 2:32 p.m., in
Room 226, Dirksen Senate Office Building, Hon. Christopher A.
Coons, Chair of the Subcommittee, presiding.
Present: Senators Coons [presiding], Klobuchar, Blumenthal,
Hirono, Welch, Tillis, and Blackburn.
OPENING STATEMENT OF HON. CHRISTOPHER A. COONS,
A U.S. SENATOR FROM THE STATE OF DELAWARE
Chair Coons. The Subcommittee of the Senate Judiciary
Committee will come to order. I'd like to thank all of our
witnesses for participating today; my colleagues for joining
me. I'd like to specifically thank Ranking Member Tillis and
his staff for working on a consensus basis to put this hearing
together and to thank Senator Blackburn and her staff, as well,
for partnering with us on this hearing.
About 10 months ago, Senator Tillis and I held a
Subcommittee hearing on artificial intelligence and copyright
law and their intersection. And I opened that hearing with the
debut of a new AI-generated song, AI, AI, a riff on Frank
Sinatra's New York, New York, with lyrics created by ChatGPT
and voice cloning technology used to mimic Frank Sinatra's
voice. The song was fun to create, with permission from the
rights holders, of course, but it also highlighted some
pressing legal questions that generative AI tools raise. Was my
song protected speech? If I hadn't gotten permission, would the
song have violated Mr. Sinatra's rights to his voice or his
style?
Since that hearing, AI-generated replicas have only grown
more pervasive, from deepfake videos of celebrities hawking
products, to songs made with voice cloning tools posing as
legitimate hits, to scam calls mimicking a panicked
grandchild's voice. AI-generated replicas of Tom Hanks and
Gayle King were used to advertise medical services, and a fake
version of Elon Musk encouraged consumers to invest in a
cryptocurrency scam. Drake, Eminem, Ariana Grande, Taylor
Swift, Beyonce are just a few of the musical artists who've
seen their voice mimicked by AI clones; and McAfee, a global
leader in online protection, found one in four American adults
have experienced an AI voice scam, with three-quarters having
lost money. Scammers using AI-generated replicas of a
grandchild's voice to trick a grandparent out of money have
become so sophisticated, both the FCC and FTC have issued
warnings.
Now, these examples all relate to commercial speech, but AI
deepfakes don't stop there. You've seen other examples:
nonconsensual explicit deepfake images and videos that is
addressed in the DEFIANCE Act that Senators Durbin and Graham
have introduced; election interference, addressed in the
Protect Elections from Deceptive AI Act that Senators
Klobuchar, Hawley, Collins, and I have introduced. Bluntly,
these issues aren't theoretical. Deepfake pornographic images
of Taylor Swift circulated broadly on X, formerly known as
Twitter, before they were taken down. A voice clone of
President Biden encouraged voters to stay home during the New
Hampshire primary. And in Slovakia, a deepfake likely had an
impact on the outcome of a national election.
In summary, as AI tools have become increasingly
sophisticated, it becomes easier to replicate and distribute
fake images of someone--fakes of their voice, fakes of their
likeness--without consent. We can't let this challenge go
unanswered, and inaction should not be an option. As President
Biden cautioned during his State of the Union, we must regulate
AI voice impersonation but do so thoughtfully, striking the
right balance between defending individual rights and fostering
AI innovation and creativity. Both Congress and the
administration have been working to strike that balance. A
bipartisan group of Senators--Young, Heinrich, Rounds, and
Schumer--convened nine AI forums last year, and Senator Schumer
has encouraged Committees to work on AI legislation on a
bipartisan basis, just as we're doing today.
That's why I was excited to release the NO FAKES Act
discussion draft last October with Senators Tillis, Blackburn,
and Klobuchar. This bill would protect people from having their
images, voices, or likenesses used to create digital replicas
that say or do things they never agreed to or would never say.
The bill accomplishes this broad goal in two ways: by holding
individuals and companies liable if they produce an
unauthorized digital replica of an individual's voice, image,
or likeness; and by holding platforms liable if they host or
distribute an unauthorized digital replica if the platform
knows the person depicted did not authorize it.
Unlike current right-of-publicity laws that many States
have enacted, which often are focused on celebrities who
monetize their likeness and leave ordinary people without a
remedy, NO FAKES Act protections would apply to all
individuals, regardless of whether they commercialize their
voices, images, or likenesses. Our bill tries to be careful to
balance these protections against free speech rights. The First
Amendment will, of course, apply to this bill, whether we say
it does or not, but we made clear, long-recognized carveouts.
Like, for example, parody and satire remain available to
creators, to continue to foster the artistic and innovative
potential of AI.
Over the past 6 months, we have had literally dozens of
meetings and received extensive feedback: hundreds, if not
thousands, of proposed revisions, tweaks, edits, wholesale
changes on the discussion draft from stakeholders who loved the
draft, who hated the draft, everyone in between. That was
exactly the point, and I appreciate the many constructive
suggestions we received. That's also the point of having a
hearing today with folks who support the bill, who question the
bill, who oppose the bill, and to have a real dialog.
Let me close. The feedback has centered around five
different sort of core technical areas: whether we should
include a notice-and-takedown structure, similar to the DMCA;
whether we've struck the right balance with First Amendment
exclusions; whether a 70-year postmortem term should be
adjusted or narrowed; whether our bill should have preemptive
impact over similar State laws; whether the bill should create
some process by which individuals with limited resources and
minimal damages can enforce their rights under the law.
So, I look forward to continuing this work with my
colleagues and, immediately following this hearing, to work to
promptly formalize the NO FAKES Act for introduction next
month. With Senator Blackburn and Senator Tillis, their great
cooperation, we have assembled a wonderful panel today with
diverse perspectives. I encourage you, as our witnesses, to
tell us what you like about the draft, what you dislike about
the draft, and be specific about what changes you would like us
to consider and why. I'll introduce the witness panel in a
moment, but let me next invite Senator Tillis to make his
opening remarks.
OPENING STATEMENT OF HON. THOM TILLIS,
A U.S. SENATOR FROM THE STATE OF NORTH CAROLINA
Senator Tillis. Thank you, Chairman Coons. As you were
going through the description of the NO FAKES Act, I was
thinking about--we've done a lot of--I love this Subcommittee,
because we actually do work here, and we actually have a bunch
of IP nerds or other interested parties showing up. We have a
lot of people interested in this space. But I really think the
NO FAKES Act is unique among the other bills that we've carried
forward, in terms of intellectual property, because it touches
everybody. You know, normally it's about patentholders or
creators, and this touches everybody, every socioeconomic
strata. It's interesting, but it's also one of the reasons why
we've got to get it right.
We've got to make sure that we come up with concrete
solutions. We don't want to overreach. There is a need for
legislation, so anyone who's in the don't fix it; it ain't
broke category, I respectfully disagree, but I'd be fascinated
to hear your testimony if we have witnesses that are of that
position. But we also don't want to miss the opportunity or we
don't want to stifle opportunities for innovation. That's why
it's so important to get it right.
We don't even know what AI is going to look like, 10 years
from now. Interestingly enough, AI is going to make AI even
more sophisticated over a much shorter period of time, so we'll
also--that's got to be instructive to our policy formulation.
But we've all seen, as Chairman Coons has indicated, replicas,
deepfakes, photos, videos, audios. We're going to show you an
example, here, shortly. And the number is just growing. So, we
have to work on it, and we have to do the fair things.
Entertainers, politicians, and the public at large have been
subject to, really, fake media for really much of the last
hundred years, but now it's getting serious, and it's producing
and multiplying at a rate that requires congressional action.
I think I want to go forward a little bit in my comments,
because I think Senator Coons did a good job of describing some
of the challenges, some of the things we want to work on with
our bill. But I'd like to--I think if staff is prepared, I want
to show you a video to give you a recent example. This was
interesting. I use AI every morning as a part of my news feed,
so I interact with generative AI for about an hour every
morning; have for about 2 years, since ChatGPT first released
their beta version of OpenAI, ChatGPT.
And it was a week or so ago that I saw the estate of Tupac
questioning a recent production. I said, more to come. Let's
study this more. And we thought it was probably interesting for
maybe folks that aren't following the issue as closely as us,
to show the video. Have we got staff ready to cue that up?
[Video is shown.]
Senator Tillis. So, that entire musical rendition is a
product of AI. And, interestingly, that image--one of those,
that name, image, and likeness, is something that is the
property of Tupac's estate. The other one is a AI-generated
image that was obviously done in violation--to the extent it
was used for commercial purposes, in violation of copyright.
So, it just gives you an example of real--this isn't a
hypothetical. This happened beginning a week or so ago, shortly
after Drake released that song. So, we've got work to do. And
legislation addressing the misuse of digital replicas will have
a multi-billion-dollar implication. We've got to get it under
control.
Now, there're a lot of questions that have to be asked, and
my office, in particular, is guilty of putting drafts out
there, knowing that they're drafts. Sometimes we even do it
sooner, without the cooperation or involvement of other
members, because we put scary stuff out there to give you a
Ghost of Christmas Future. In this case, we didn't do that. We
tried to work on putting out a discussion draft that makes
sense, but we've got a lot of things that we have to work out.
You know, the questions that we need answered: Is it wise to
mandate that individuals have no right to license out of their
digital likeness unless they retain counsel? Should we create
an exception for harmless, noncommercial users? Should there be
a notice-and-takedown provision? I mean, there's a lot. There's
a litany. I'm not going to go through all of them. I hope that
you all can come up with other ones. But I'll submit the rest
of my written statement for the record.
[The information appears as a submission for the record.]
Senator Tillis. But we have to act. Hopefully, in this
Congress we can act, which means we have to move very, very
quickly or, at a minimum, lay down a baseline that we can pick
up, when we come back with a new Congress, and get it right.
So, I look forward in advance to everybody's active
collaboration. I will always give you the same warning I give
everyone: The only thing that really makes me mad is when I see
somebody trying to, through guerilla warfare, undermine the
good faith efforts of this Committee and my colleague. If
you're at the table, you can have an influence. If you're not
at the table, you're going to be on the table.
So, why don't everybody just recognize our office is open
to constructive criticism? Use cases of where the policy
doesn't make sense. But if you're in the category of, it ain't
broke; don't fix it, you're not up with modern times. And I
look forward to a good, productive hearing today, and thank you
in advance for your productive collaboration as this
legislation moves forward.
Chair Coons. Thank you, Senator Tillis. And thank you for
another positive and engaging hearing. It's been a great
experience serving on this Committee with you.
Today we welcome six witnesses to testify about the NO
FAKES Act. Our first witness is Robert Kyncl, CEO of Warner
Music Group, who has a lot of experience. He spent over a
decade as YouTube's chief business officer, among a number of
other business engagements. Next we have twigs--thank you,
twigs, for joining us today--a singer, songwriter, producer,
dancer, and actor who has used AI to help her create and also
has had personal experience with unauthorized AI deepfakes.
It's great to have a voice present from the creative community.
Next we have Duncan Crabtree-Ireland, national executive
director and chief negotiator for SAG-AFTRA, the Screen Actors
Guild-American Federation of Television and Radio Artists, a
labor union representing 160,000 members who work in film,
television, music, and more. Also the voice of the creative
community. Then we have Ben Sheffner, senior vice president,
associate general counsel for law and policy at the Motion
Picture Association, where he specializes in copyright and
First Amendment law. Thank you, Ben.
We welcome Graham Davies, presidency of the Digital Media
Association, an organization that represents principally audio
streaming companies, platforms like Spotify and YouTube. Mr.
Davies also has a history as a musician and songwriter
advocate. Finally, we'll hear from Lisa P. Ramsey, professor of
law at the University of San Diego School of Law, where she
teaches and writes on the intersection of free speech rights
and intellectual property law.
After I swear in the witnesses, each will have 5 minutes to
provide a summary of your opening statement. The Senators have
your written statements. Then we'll proceed to questioning.
Each Senator gets 5 minutes for the first round. We will likely
have two or even three rounds of questioning, time and
attendance permitting. Witnesses, would you please stand? Raise
your right hand to be sworn in.
[Witnesses are sworn in.]
Chair Coons. Thank you all. Let the record reflect the
witnesses have been sworn. Mr. Kyncl, you may proceed with your
opening statement.
STATEMENT OF ROBERT KYNCL, CHIEF EXECUTIVE OFFICER, WARNER
MUSIC GROUP, NEW YORK, NEW YORK
Mr. Kyncl. Chairman Coons, Ranking Member Tillis, and
Members of the Subcommittee, I'm Robert Kyncl, chief executive
officer of the Warner Music Group. Being here today is
something I could not have imagined as a young boy growing up
behind the Iron Curtain in communist Czechoslovakia. In 1992, I
crossed the Atlantic and attended State University in New York,
and there I met an amazing woman from the Dominican Republic
who eventually became my wife, and now we have two amazing
American daughters. I'm a proud U.S. citizen, and I have deep
appreciation for the freedoms at the heart of this great
country, having grown up without them.
For the past 25 years, I've been a tech and media
executive. I joined Warner Music last year after 12 years at
YouTube and 8 years at Netflix. Warner Music is home to an
incredible array of artists and songwriters who are moving
culture across the globe. One of those artists, FKA twigs, is
here with me today, and she's an extraordinarily gifted singer,
songwriter, actor, and performer. I would like to acknowledge
and thank Duncan Crabtree-Ireland, who led the recent
collective bargaining agreement negotiations between SAG-AFTRA
and record labels that addresses concerns regarding AI and
defends artists' rights.
Music has so often been the canary in the coal mine for the
broader trends in our society. More than any other form of
communication or entertainment, music drives culture and
innovation, and that's happening again with generative AI.
Today, music companies are helping artists, rights holders, and
tech companies figure out this new world, which is both
exciting and daunting. It's our job to not only help amplify
artists' creativity but to protect their rights, their
livelihoods, and their identities.
Across the industry, legends from Roberta Flack to the
Beatles have embraced AI as a tool to enhance their creativity.
At the same time, generative AI is appropriating artists'
identities and producing deepfakes that depict people doing,
saying, or singing things that have never happened before. My
accent is a vestige to my Eastern European upbringing. You can
hear my identity in my voice. Through AI, it is very easy for
someone to impersonate me and cause all manner of havoc. They
could speak to an artist in a way that could destroy our
relationship. They could say untrue things about our publicly
traded company to the media that would damage our business.
Unfettered deepfake technology has the potential to impact
everyone, even all of you. Your identities could be
appropriated and used to mislead your constituents. The truth
is, everyone is vulnerable: families defrauded by voice clones
pretending to be relatives, people placed in pornography
without their consent, schoolchildren having their faces
inserted into humiliating scenes. Some people have spoken of
responsible AI as a threat to freedom of speech, but it's
precisely the opposite. AI can put words in your mouth, and AI
can make you say things you didn't say or don't believe. That's
not freedom of speech.
We appreciate the efforts of this Committee to address this
problem, including the NO FAKES Act discussion draft authored
by Chairman Coons, Ranking Member Tillis, Senator Blackburn,
and Senator Klobuchar. Your leadership kickstarted efforts in
this area, and we strongly support the bipartisan No AI FRAUD
Act introduced in the House earlier this year by
Representatives Salazar and Dean and the recently enacted ELVIS
Act in Tennessee.
As the Committee moves forward the introduction of a Senate
bill, there are three elements that those should contain, to be
effective. One, an enforceable intellectual property right for
likeness and voice. Each person should be allowed to license or
deny that right on free-market terms and seek redress for
unauthorized uses. Two, respect for an important First
Amendment principles, without going any further and providing
loopholes that create more victims. And, three, effective
deterrents. To incentivize a vibrant and commercial
marketplace, we need to maintain consequences for AI model
builders and digital platforms that knowingly violate a
person's property rights.
I applaud the Committee for its leadership in addressing
these challenging and rapidly developing issues with urgency.
Congress should pass legislation this year, before the genie is
out of the bottle, while we still have a chance to get this
right. I look forward to answering your questions. Thank you.
[The prepared statement of Mr. Kyncl appears as a
submission for the record.
Chair Coons. Thank you, Mr. Kyncl. Twigs.
STATEMENT OF TAHLIAH DEBRETT BARNETT (``FKA TWIGS''),
SINGER, SONGWRITER, PRODUCER, DANCER, AND ACTOR, LONDON, UNITED KINGDOM
Ms. Barnett. As artists, we dedicate a lifetime of hard
work and sacrifice in the pursuit of excellence, not only in
expectation of achieving commercial success and critical
acclaim but also in the hope of creating a body of work and
reputation that is our legacy. So, why am I here today? I am
here because my music, my dancing, my acting, the way my body
moves in front of the camera, and the way that my voice
resonates for a microphone is not by chance. They are essential
reflections of who I am. My art is a canvas on which I paint my
identity and the sustaining foundation of my livelihood.
It is the very essence of my being, yet this is under
threat. AI cannot replicate the depth of my life journey, yet
those who control it hold the power to mimic the likeness of my
art, replicate it, and falsely claim my identity and
intellectual property. This prospect threatens to rewrite and
unravel the fabric of my very existence. We must enact
regulation now to safeguard our authenticity and protect
against misappropriation of our inalienable rights.
Three decades ago, we did not realize that the internet
would embed itself so deeply into the core of our everyday
lives. Policies and controls to keep pace with the emergence of
the technology were not put in place to protect artists, young
people, and those that are vulnerable, and it ran away with us.
AI is the biggest leap in technological advancement since the
internet. You know the saying, fool me once, shame on you; fool
me twice, shame on me? If we make the same mistake with the
emergence of AI, it will be shame on us.
Let me be clear. I am not against AI. As a future-facing
artist, new technologies are an exciting tool that can be used
to express deeper emotions, create fantasy worlds, and touch
the hearts of many people. In the past year, I have developed
my own deepfake version of myself that is not only trained in
my personality but that can also use my exact tone of voice to
speak many languages. These and similar emerging technologies
are highly valuable tools. This, however, is all under my
control, and I can grant or refuse consent in a way that is
meaningful.
What is not acceptable is when my art and my identity can
simply be taken by a third party and exploited falsely for
their own gain, without my consent, due to the absence of
appropriate legislative control and restriction. History has
shown us time again that in moments of great technological
advancement, those in the arts have been the first to have
their work exploited and, in many instances, fraudulently
commoditized. Soon after, it follows that the general and more
vulnerable public suffer the same types of image-and voice-
related exploitation. By protecting artists with legislation at
such a momentous time in history, we are protecting a 5-year-
old child, in the future, from having their voice, likeness,
and identity taken and used as a commodity without prior
consent.
I stand before you today because you have it in your power
to protect artists and their work from the dangers of
exploitation and the theft inherent in this technology if it
remains unchecked. I am here on behalf of all creators whose
careers depend deeply on the ability to create, safe in the
knowledge that they can maintain tight control over their own
art, image, voice, and identity. Our careers and our
livelihoods are in jeopardy, and so potentially are the wider
image-related rights of others in society. You have the power
to change this and safeguard our future.
As artists and, more importantly, human beings, we are a
facet of our given, learned, and developed identity. Our
creativity is a product of this lived experience overlaid with
years of dedication to qualification, training, hard work,
and--dare I say it--significant financial investment and
sacrifice. That the very essence of our being, at its most
human level, can be violated by the unscrupulous use of AI to
create digital facsimile that purports to be us and our work is
inherently wrong. It is therefore vital that, as an industry
and as legislators, we work together to ensure we do all we can
to protect and create an intellectual rights system as well as
protect the very basis of who we are.
We must get this right. You must get this right, before
it's too late. Thank you.
[The prepared statement of Ms. Barnett appears as a
submission for the record.
Chair Coons. Thank you. Mr. Crabtree-Ireland.
STATEMENT OF DUNCAN CRABTREE-IRELAND, NATIONAL
EXECUTIVE DIRECTOR AND CHIEF NEGOTIATOR, SCREEN
ACTORS GUILD-AMERICAN FEDERATION OF TELEVISION
AND RADIO ARTS, LOS ANGELES, CALIFORNIA
Mr. Crabtree-Ireland. Thank you very much, Chairman Coons,
Ranking Member Tillis, and the Members of the Subcommittee on
Intellectual Property. My name is Duncan Crabtree-Ireland. I'm
the national executive director of SAG-AFTRA, the country's
largest labor union for entertainment and media artists. And
I'm here today to testify in support of the NO FAKES Act. Our
members believe that AI technology, left unregulated, poses an
existential threat to their ability to, one, require consent
for the creative use of their digital representation; two,
receive fair payment for use of their voice and likeness; and,
three, to protect against having to compete against themselves,
their own digital self, in the marketplace.
I'm the chief negotiator for the union's contracts,
including last year's historic agreement with the major
entertainment studios, which was only finalized after the
longest entertainment industry strike in over 40 years--a
strike that lasted nearly 4 months. The strikes and the
public's response to them highlighted that the entertainment
industry and the broader public understand that AI poses real
threats to them, and they fully support protections against
those threats. For an artist, their image and likeness are the
foundations of their performance, brand, and identity developed
over time through investment and hard work. SAG-AFTRA has long
fought for right-of-publicity laws and voice and image
protections. The exponential proliferation of artificial
intelligence technologies--technologies which allow for rapid
and realistic fakes of voices and likenesses in audiovisual
works and sound recordings--makes this fight urgent for our
members.
Enshrining this protection as a Federal intellectual
property right will ensure our members, creative artists, and,
frankly, all of us are protected and service providers provide
the same protections to individuals' images, likenesses, and
voices that they provide now for other intellectual property
rights. These rights should be transferable and descendible,
just like any other intellectual property or any kind of
property someone owns, with durational limitations on transfers
during one's lifetime, to ensure that we don't enter into an
era of digital indentured servitude, just as actress and SAG-
AFTRA member Olivia de Havilland fought to establish the 7-year
rule to end long-term abusive contracts in the old studio
system.
Some would argue that there should be broad, categorical
First Amendment-based exemptions to any legislation protecting
these important rights. There are no stronger advocates for the
First Amendment than our members. They rely on their First
Amendment rights to tell the stories that artists in other
countries are often too endangered to tell. However, the
Supreme Court has made clear over a half a century ago that the
First Amendment does not require that the speech of the press
or any other media, for that matter, be privileged over
protection of the individual being depicted. To the contrary,
the courts apply balancing tests which determine which right
will prevail.
These balancing tests are critical, and they are
incorporated into the discussion draft of the NO FAKES Act.
They ensure that the depicted individual is protected and
rewarded for the time and effort put into cultivating their
persona while not unduly burdening the right of the press to
report on matters of public interest or the entertainment media
to tell stories. At the same time, these tests help ensure the
depicted individual is not compelled to speak for the benefit
of third parties who would misappropriate the value associated
with the persona they have carefully crafted. With new AI
technologies that can now realistically depict an individual's
voice or likeness with just a few seconds of audio or even a
single photograph, and with constantly evolving capabilities of
these technologies, it is even more important that broad
categorical exemptions be avoided and that the courts be
empowered to balance the competing interests.
It's also essential that action be taken to address these
harms now. Our members, the public, and our society are being
impacted right now by the abuse of deepfake technology, and we
must take timely action. Just as one of many examples of the
abusive deepfake technology, during the ratification campaign
for our contract after the strike last year, an unknown party
on the internet created an unauthorized deepfake video of me
saying false things about our contract and urging members to
vote against it. Anathema to me, as someone who devoted my life
for more than a year to a contract I deeply believe in.
There was no Federal right to protect me, no takedown
right, and tens of thousands of people were misled about
something that really mattered to so many of us. It's neither
necessary nor appropriate to wait for broader artificial
intelligence regulation to be adopted. This narrow and
technology-neutral approach can and should proceed
expeditiously forward. The companies behind many of these
technologies are asking for rules so they better understand the
appropriate boundaries on their conduct. The NO FAKES Act will
provide them with important guidance, while helping to ensure
individuals are protected from exploitation that puts their
livelihoods and reputations at risk. Thank you again for this
opportunity to speak today, and I look forward to answering
your questions.
[The prepared statement of Mr. Crabtree-Ireland appears as
a submission for the record.
Chair Coons. Thank you, Mr. Crabtree-Ireland. Mr. Sheffner.
STATEMENT OF BEN SHEFFNER, SENIOR VICE PRESIDENT
AND ASSOCIATE GENERAL COUNSEL, LAW AND POLICY,
MOTION PICTURE ASSOCIATION, INC.,
LOS ANGELES, CALIFORNIA
Mr. Sheffner. Chair Coons, Ranking Member Tillis, Members
of the Subcommittee, thank you for the opportunity to testify
today on behalf of the Motion Picture Association and our
member studios about legislation to regulate the use of digital
replicas. For over a century, the MPA's members have employed
innovative new technologies to tell compelling stories to
audiences worldwide. From the introduction of recorded sound in
the 1920's, color in the 1930's, the dazzling special effects
in movies like this year's Dune: Part Two, the MPA's members
have long used technology to allow filmmakers to bring their
vision to the screen in the most compelling way possible.
Artificial intelligence is the latest such innovation impacting
our industry.
MPA sees great promise in AI as a way to enhance the
filmmaking process and provide an even more compelling
experience for audiences, but we also share the concerns of
actors and recording artists about how AI can facilitate the
unauthorized replication of their likenesses or voices to
supplant performances by them, which could potentially
undermine their ability to earn a living practicing their
craft. The NO FAKES Act is a thoughtful contribution to the
debate about how to establish guardrails against abuses of such
technology; however, legislating in this area necessarily
involves doing something that the First Amendment sharply
limits: regulating the content of speech.
It will take very careful drafting to accomplish the bill's
goals without inadvertently chilling or even prohibiting
legitimate, constitutionally protected uses of technology to
enhance storytelling. I want to emphasize, this is technology
that has entirely legitimate uses, uses that are fully
protected by the First Amendment and do not require the consent
of those being depicted.
Take the classic 1994 film, Forrest Gump, which depicted
the fictional Forrest character, played by Tom Hanks,
navigating American life from the 1950's through the 1980's,
including by interacting with real people from that era.
Famously, the filmmakers, using digital replica technology
available at the time, had Forrest interact and even converse
with Presidents--or, should I say, former Senators--Kennedy,
Johnson, and Nixon. To be clear, those depictions did not
require the consent of their heirs. And requiring such consent
would effectively grant heirs or their corporate successors the
ability to censor portrayals they don't like, which would
violate the First Amendment.
In my written testimony, I detail some specific suggestions
we have for improving the NO FAKES draft so that it addresses
real harms without encroaching on First Amendment rights. Here,
I'll highlight just four points. First, getting the statutory
exemptions right is crucial. And I want to thank the drafters
for getting much of the way there. Those exemptions give
filmmakers the clarity and certainty they need to determine
whether to move forward with spending tens, even hundreds, of
millions of dollars on a movie or TV series. If the statutory
exemptions are not adequate, some producers will simply not
proceed with their projects, a classic chilling effect that the
First Amendment does not allow.
Second, the bill should preempt State laws that regulate
the use of digital replicas in expressive works. Simply adding
a Federal layer on top of the existing patchwork of State laws
would only exacerbate the problems associated with inconsistent
laws in this area.
Third, the scope of the right should focus on the
replacement of performances by living performers. Going beyond
that risks sweeping in wide swaths of First Amendment-protected
speech which would make the statute vulnerable to being struck
down on overbreadth grounds.
And, fourth, the definition of digital replica must be
focused on highly realistic depictions of individuals. It
should not encompass, for example, cartoon versions of people
you might see on shows like The Simpsons or South Park.
And, last, before legislating, MPA urges the Subcommittee
to first pause and ask whether the harms it seeks to address
are already covered by existing law, such as defamation, fraud,
the Lanham Act, or State right-of-publicity law. Often, the
answer will be yes, indicating that a new law is not necessary.
And if there is indeed a gap in the law--for example, regarding
pornographic or election-related deepfakes--the best solution
is narrow, specific legislation targeting that specific
problem. Thank you again for the opportunity to testify today,
and I welcome your questions.
[The prepared statement of Mr. Sheffner appears as a
submission for the record.
Chair Coons. Thank you, Mr. Sheffner. Mr. Davies.
STATEMENT OF GRAHAM DAVIES, PRESIDENT AND CHIEF
EXECUTIVE OFFICER, DIGITAL MEDIA ASSOCIATION,
WASHINGTON, DC
Mr. Davies. Good afternoon, Chairman Coons, Ranking Member
Tillis, and thank you to the Committee for giving me the
opportunity to speak to you today on this important issue. My
name is Graham Davies, and I'm president and CEO of the Digital
Media Association, representing the leading music streaming
services. We support the Committee's efforts to bring forward
legislation at the Federal level, which should preempt existing
State laws, to keep pace with new technology. We join you in
the objective of ensuring there are appropriate protections for
individuals' likenesses. This is an important issue for us all,
and DiMA supports efforts to develop a clear and balanced way
forward.
Our members benefit from clarity in law and providing fans
with great experiences. Indeed, our members have a strong track
record of licensing complex rights to deliver music to fans.
They work closely with record labels and music publishers, with
whom they have long relationships and robust contracts. This is
our common objective. But any new or increased rights should be
appropriate and targeted. They should not come at the expense
of important freedoms of speech or creative expression, nor
should they be overly broad to the point of creating confusion
or needless litigation, over the true objective of protecting
personhood.
The NO FAKES Act proposes to sweep in a broad range of
legitimate replicas and downstream activities within its scope.
The current draft punishes good and bad actors alike. Any new
rights should not undermine the global content supply chains on
which the streaming industry depends. We are still in the early
stages of the application of AI by the artistic community, but
we see that the existing practices for taking down illegal or
deceptive content continue to suffice in this new context.
Streaming services are the last point in the supply chain.
Only the originator of the content and the label who delivers
it to the services have the information necessary to establish
whether the content is legitimate or not. Streaming services do
not have any way to know the complex chain of rights behind the
content they receive from labels and distributors. To address
the harms caused when AI technology is used to imitate a
musical artist, a celebrity, or other public figure, we believe
the Committee's objectives would be best achieved if new
legislation was developed from the existing right-of-publicity
laws.
This would have a number of advantages. Firstly, there is a
body of existing case law on how First Amendment protections
can be balanced with the individual's right-of- publicity.
Second, liability sits squarely with the bad actors: those who
create the deceptive content and first place it into the public
sphere. And, thirdly, the focus is on commercial use, with
actual damages, which we believe are proven to be a sufficient
deterrent.
Establishing a Federal law that preempts the existing
patchwork of privacy and publicity laws is both beneficial and
necessary. Music streaming is a global industry. We believe
that rights pertaining to the person should remain inextricably
tied to the individual for the duration of their life. This
ensures that each person is always able to retain control over
how their voice is used. The discussion draft released by Chair
Coons, Ranking Member Tillis, and Senators Blackburn and
Klobuchar has been helpful to foster dialog and encourage all
stakeholders to think about these complex issues. I've included
more in my written testimony which is intended to support the
next stages of the discussion, and DiMA looks forward to
continued work with the Committee. Thank you.
[The prepared statement of Mr. Davies appears as a
submission for the record.
Chair Coons. Thank you, Mr. Davies. Professor Ramsey.
STATEMENT OF LISA P. RAMSEY, PROFESSOR OF LAW,
UNIVERSITY OF SAN DIEGO SCHOOL OF LAW,
SAN DIEGO, CALIFORNIA
Professor Ramsey. Chair Coons, Ranking Member Tillis, and
other Members of the Subcommittee, thank you for the
opportunity to testify today about the First Amendment
implications of the proposed NO FAKES Act. I'm a professor of
law at the University of San Diego School of Law. I teach
intellectual property classes at USD, and my scholarship
focuses on the potential conflicts between trademark laws and
the right to freedom of expression.
The First Amendment of the U.S. Constitution commands that
Congress shall make no law that abridges the freedom of speech.
Congress generally lacks the power to restrict expression
because of its message, ideas, subject matter, or content. This
rule is subject to a few limited exceptions for historically
unprotected speech, such as fraudulent speech and obscenity,
but content-based regulations of speech are generally presumed
invalid unless the government can prove the law is
constitutional.
The NO FAKES Act imposes restrictions on the content of
speech. It targets the harms caused by the unauthorized
creation and dissemination of digital replicas or deepfakes of
individuals in recordings that are nearly indistinguishable
from that person's actual voice, image, or visual likeness.
When the Act applies to the use of digital replicas to
impersonate people in fraudulent speech or misleading
commercial speech, it is consistent with the First Amendment.
There's also no conflict with the First Amendment when the Act
restricts the use of digital replicas in sexually explicit
deepfakes without consent if those images or videos constitute
obscene speech or child pornography.
The problem is that the current version of the NO FAKES Act
also regulates nonmisleading speech that is protected by the
First Amendment. Congress must therefore prove that the Act
satisfies constitutional scrutiny. The law must be narrowly
tailored to directly and materially further its goals and not
harm speech that's protected by the First Amendment more than
necessary. Strict scrutiny analysis may be required when the
government is regulating the unauthorized use of digital
replicas in political messages, news reporting, entertainment,
and other types of noncommercial speech that's fully protected
by the First Amendment.
As it is currently drafted, I believe the NO FAKES Act is
not consistent with the First Amendment because the law is
overbroad and vague; however, I think a revised version of the
law could satisfy intermediate and strict constitutional
scrutiny. There are three ways that Congress can better protect
First Amendment interests in the law. First, it's critical that
the law not suppress or chill protected speech more than
necessary. The Senate's proposed NO FAKES Act does a better job
than the No AI FRAUD Act in setting forth specific exemptions
from liability for certain nonconfusing uses of another's
image, voice, or likeness. The law can still be improved in
certain ways that I discuss in my written testimony.
It is also important that Congress not enact a strict
liability rule for online service providers that host
expression covered by the NO FAKES Act. Specific and actual
knowledge of the direct infringer's use of an unauthorized
digital replica should be required for liability. Online
service providers should implement a notice-and-takedown
system, to make it easier to remove unauthorized deepfakes that
violate the law. Accused infringers must also, though, be able
to challenge takedown requests by filing a counter-notification
with the platform.
My second recommendation is for Congress to create separate
causes of action that target the different harms caused by
unauthorized uses of digital replicas. This includes, number
one, the use of deepfakes to impersonate individuals in a
deceptive manner; number two, uses in sexually explicit
deepfakes; and, number three, uses that substitute for an
individual's performance that they typically would have created
in real life, such as a performance in a song or movie. These
causes of action should have different requirements and
distinct speech-protective exceptions.
My third recommendation is that Congress ensure each
provision of the law adequately protects speech interests.
Congress can better protect expressive values by allowing the
new Federal statute to preempt the inconsistent State laws that
protect the right-of-publicity and digital replica laws that
restrict the unauthorized use of digital replicas. If licensing
of digital replica rights is allowed by the Act, individuals
should be able to consent for each different use of their
digital replica.
Allowing others to control a person's identity rights
through a broad licensing agreement will work at cross purposes
with many of the stated goals of this proposed legislation. It
could potentially lead to greater AI-generated deception of the
public. It can also stifle the right of people to make a living
through their performances and result in the use of their image
or their voice in sexually explicit material that was
authorized by the broad terms of a licensing agreement.
I encourage Congress to continue to protect the interests
of both public figures and ordinary people, in the NO FAKES
Act, and I also encourage you to continue consulting with
stakeholders, academics, and attorneys with expertise in this
field of law. I look forward to answering your questions as you
continue to improve the Act. Thank you.
[The prepared statement of Professor Ramsey appears as a
submission for the record.
Chair Coons. Thank you. Thank you to all six of our
witnesses for your preparation, your engagement. I'm going to
start with some questions about exploring how AI replicas are
impacting individuals and entertainment businesses and then use
a subsequent round to get into your perspectives on specific
potential revisions to the NO FAKES Act.
Mr. Crabtree-Ireland, thank you for sharing your personal
experience of an AI-generated deepfake. This was in the context
of the ratification fight for the most recent contract. Given
your experience, should a digital replica right apply to all
individuals, regardless of whether they're commercializing
their image, voice, or visual likeness? You primarily represent
people who make a living, who commercialize their image, voice,
or visual likeness. Why should we have this available to
everyone?
Mr. Crabtree-Ireland. No, it's a great question, Chairman.
I think, yes, we support a right that's available to everyone.
Obviously, twigs, myself, others, Mr. Kyncl, have explained the
impact that this can have on people who make a living and whose
career is based on their image, likeness, or voice, but the
impacts are so obvious and so real for so many Americans,
outside of the scope of just a commercialized use. And the
example I gave, in my mind, is not a commercial use example.
This is an example that could apply to anyone, and the impact
is so serious. So, yes, we do support this right on a broader
basis, and it should be applicable to everyone.
Chair Coons. Twigs, could you help us understand how you're
using AI as a creative tool, on the one hand, and then briefly
tell us a little bit about your experience with AI deepfakes
and what you think the future of your industry looks like if we
don't heed your urgent call for us to act?
Ms. Barnett. Well, over the past year, I've been creating
an AI version of myself that can use my tone of voice exactly
to speak in multiple languages. I've done this to be able to
reach more of my fans and to be able to speak to them in the
nuance of their language. So, I've currently explored French,
Korean, and Japanese, which is really exciting for me. It means
that, even with my upcoming album, I can really explain in
depth what it's about, creatively. It also allows me to spend
more time making art. Often, being a music artist or any artist
in this day and age requires a lot of press and a lot of promo,
a lot of one-liners. So, it means if it's something simple that
doesn't really require my heart, I can do a one-liner and give
it to people, to promote a piece of work, and, you know, it's
harmless, but ultimately I can spend more time making something
that's really meaningful for my fans. And the next question--
oh, you asked how----
Chair Coons. Your own----
Ms. Barnett [continuing]. Have I been----
Chair Coons [continuing]. Experience with deep----
Ms. Barnett. Yes. So, there are songs online,
collaborations with myself and other artists, that I didn't
make. It makes me feel vulnerable, because first of all, as an
artist, I think the thing that I love about what I do is that
I'm very precise. I take my time with things, and it's really
what I'm very--I'm very proud of my work, and I'm very proud of
the fact that I think my fans really trust me, because they
know that I just put so much deep meaning of my north star into
what I do.
So, the fact that somebody could take my voice, change
lyrics, change messaging, maybe work with an artist that I
didn't want to work with or maybe work with an artist that I
wanted to work with and now the surprise is ruined--it really
leaves me very raw and very vulnerable. I think that if
legislation isn't put in place to protect artists, not only
will we let artists down who really care about what we do, that
have spent a long time developing themselves, developing the
way that we work, but it also would mean that fans wouldn't be
able to trust people that they've spent so many years investing
in.
It would affect us spiritually. It would affect us
financially. And it makes--honestly, if I'm honest with you,
I'm just surprised that we're even having this conversation,
because it feels so painfully obvious to me that it's hard to
even find the language, if I'm completely honest with you.
Chair Coons. There're a lot of things that are painfully
obvious that call out for Congress to act----
[Laughter.]
Chair Coons. So, your surprise is not unusual.
[Laughter.]
Chair Coons. It is broadly spread.
Ms. Barnett. But ultimately what it boils down to is my
spirit, my artist, and my brand is my brand, and I've spent
years developing it, and it's mine. It doesn't belong to
anybody else, to be used in a commercial sense or cultural
sense or even just for a laugh. You know? I am me. I am a human
being. And we have to protect that.
Chair Coons. Thank you. Mr. Kyncl, if I might, just
briefly--we've seen a steady increase in the quality of
deepfakes, with songs on streaming platforms that are virtually
indistinguishable from talented artists like twigs. What are
the challenges that AI deepfakes are creating, sort of long
term, for both the music business and for fans, as well as for
performers?
Mr. Kyncl. So, I think twigs addressed one of those, and no
one can do that better than what she just did. I think the
second one is that, when you have these deepfakes out there,
the artists are actually competing with themselves for revenue
on streaming platforms, because there's a fixed amount of
revenue within each of the streaming platforms, and if
somebody's uploading fake songs of twigs and those songs are
eating into that revenue pool, there's less left for her
authentic songs. So, that's the economic impact of it, long
term. And the volume of content that will then flow into the
digital service providers will increase exponentially, which
will be harder for the artist to be heard and to actually reach
lots of fans. So, creativity, over time, will be stifled.
Chair Coons. So, as you both put it, there's both a
relationship impact, a spiritual impact, a financial impact, a
broader ecosystem of creativity impact. Senator Tillis, I turn
to you.
Senator Tillis. Thank you, Chairman Coons. And, again,
thank you all for being here. Ms. Ramsey, I'm going to start
with you and then have others who may have an opinion on it.
You mentioned notice-and-takedown in your comments. This is a
strict liability bill, in its current form. Some of us think
that maybe we have to weigh into that. You also talked a bit
about having, I guess, that--having the individual who's been
informed of takedown having some recourse. Can you talk a
little bit more about that, briefly?
Professor Ramsey. Sure. You might have a situation where
somebody challenges your own personal use of your identity
online, and they're the one that's the bad actor, but they file
a complaint with the online service provider, and the online
service provider, who wants to avoid liability, automatically
takes it down. That's one possibility.
Another would be that the person who is disseminating this
image actually has a defense, right? An exception applies to
this particular use. It might be news reporting or parody. And
so it's critical for the online service provider to be able to
put that expression back up, if it actually does not violate
the law. Under the copyright laws, my understanding is that
once the information's put back up, it stays up unless the
copyright owner files a lawsuit.
So, what's great about the takedown-and-notice procedure is
that it allows ordinary people to get these unauthorized uses
off the internet. And I think that's one real benefit of having
a notice-and-takedown procedure and encouraging companies to
adopt one. There are some challenges, though, with notice-and-
takedown procedures that folks like Eric Goldman and others
have talked about, so it's great that you're talking to
interested parties when you figure out these issues.
Senator Tillis. Anybody here have an opinion counter to
that? Okay. Mr. Kyncl, can you walk me through what artist,
what rights typically are granted to record labels under
exclusive sound recording agreement, and are likenesses
included in that?
Mr. Kyncl. So, it's a pretty wide range of rights, anywhere
from a full copyright right to distribution-only rights, where
the copyright remains with the artist. And increasingly so,
they include likeness, as well, because we're--as you can
imagine, as we work on open platforms with lots of user-
generated content, we are the ones who have a staff of people
that is working to issue notices, claim the content, take down
the content. And, increasingly, we need the name, image,
likeness, and voice rights in order to actually act on that, on
the artist's behalf, with the platforms.
Senator Tillis. I think you believe that new digital
replica rights need to be fully transferable?
Mr. Kyncl. Yes.
Senator Tillis. Why isn't a license enough?
Mr. Kyncl. I think it should be at the choice of the
artist. The artist should have a choice to either transfer or
license.
Senator Tillis. Okay. Mr. Sheffner, State-level right-of-
publicity laws restricting commercial speech have existed for
many decades. They've developed their own case law, and they're
well understood. The new digital replica right proposed by the
NO FAKES Act would affect noncommercial speech beyond what most
State laws currently cover. Can you explain how novel this
proposed right would be in the context of existing right-of-
publicity laws, and how should we consider preempting similar
State-level digital replica laws, especially when it's such new
territory?
Mr. Sheffner. So, thank you for that question, Senator
Tillis. And you're absolutely right that most State right-of-
publicity laws, which have existed for more than a century, are
really limited to commercial uses. That's in advertisements or
on merchandise. What Congress is considering doing here is
really novel. Although it's sometimes described as right-of-
publicity, we think it's fundamentally different in that it
would apply in expressive works like movies, TV shows, songs
which are fully protected by the First Amendment.
So, there have developed a robust body of case law in the
traditional commercial right-of-publicity context which says,
yes, it applies if you put somebody's face on a billboard or
use it in an advertisement, on a lunchbox, but it doesn't apply
if, for example, you're making a biopic or a docudrama about
somebody. You can't use right-of- publicity law to censor those
portrayals. Again, this is a novel form of right which is going
to be subject to heightened constitutional scrutiny, like
Professor Ramsey described.
Because it applies in expressive works, it's really
important, up front, to provide some clarity to film producers
so that when they're about to embark on a project, they know
what's allowed and what's not. And if it's too vague, if it's
too uncertain, they're going to shy away from using this
technology to engage in those sorts of portrayals. Again, that
chills speech, and the First Amendment case law says that a
statute is vulnerable to being struck down if it chills
constitutionally protected speech.
Senator Tillis. Yes, which is why we absolutely have to get
it right. I think there's general consensus that we have to
make progress, but the challenges of all this work being struck
down are significant. We have to do the legwork. Thank you.
I'll have a second round.
Chair Coons. All right. Thank you, Senator Tillis. Senator
Hirono.
Senator Hirono. Thank you, Mr. Chairman and Ranking Member
Tillis, for bringing this bill before us. And as you say, Mr.
Chairman, the bill has gone through a lot of input from a lot
of different groups. And if I listened to your testimony
accurately, it doesn't sound as though any of you think that we
should not do something that will protect--I like the framing
of protecting personhood. Any of you think that we don't need
to do anything in this area? Okay.
So, looking at this statute, then, why don't we go down the
list very quickly? What do you like most about the current
bill--and we'll just start with Mr. Kyncl--about the current
bill, and what is the most important thing you would want to
change, if anything? And if you can just keep your answer
really short.
Mr. Kyncl. I'll start with what I believe it needs to
contain, which is it needs to contain consent for the use of
people's name, likeness, and voice, to train AI models and
create outputs. That needs to happen. Second, it needs to
contain monetization, which is fair market license that that
person can exercise, you know, through that consent. But in
order for that to happen, and in order for all of that to be
operationalized by the platforms, we need two things: one,
which is provenance of the content that generative AI models
are trained on and that they're outputting to be retained,
which means they should keep sufficiently detailed records on
what they trained on, so that later on that provenance can be
embedded into watermarks which are recognized by the platforms
on which the content is.
Senator Hirono. I mean, the point is, it sounds like,
consent is the critical part of this. Consent of the creator--
--
Mr. Kyncl. And the provenance----
Senator Hirono [continuing]. Of the image.
Mr. Kyncl [continuing]. Of the content.
Senator Hirono. Okay.
Mr. Kyncl. And we are good at tracing provenance on luxury
clothing, on cheese, on wine. We should be able to do it on
intellectual property, as well.
Senator Hirono. Going down the line. We're talking about
this particular bill. Is there something in this bill that you
think is the most critical aspect of the bill, that you
support? And is there anything there you would change in the
bill?
Ms. Barnett. I think the most important thing is to put the
power in the hands of the artists. I want to be in control of
my likeness, my brand, my legacy. I have sacrificed so many
years to be good at dancing, at singing; so much financial
input, so much time. And I do it in the name of my legacy. I do
it so that one day I can look back at my body of work and say,
that was me. And that's what I want to be protected in the
bill.
Mr. Crabtree-Ireland. Thank you, Senator. I think what I
like most about this bill is the fact that it is broader than
limiting it to commercial use. The fact is the commercial use
limitation may have worked 100 years ago. A commercial use
limitation does not solve the problems that we face today,
especially because of generative AI. And we need the breadth
that is reflected in this legislation.
I think in terms of--if there were one thing that I would
change in it, I would adopt a durational limitation on transfer
or even licenses of these rights during lifetime. I think may
not be as necessary after death, but during a lifetime, I think
it's essential in order to make sure that someone doesn't
improvidently grant a transfer of rights early in their
lifetime that turns out to be unfair to them, and I think
there's various standards we could look at for an appropriate
duration.
Senator Hirono. So, 70 years is a bit long.
Mr. Crabtree-Ireland. Well, I'm sorry, the 70 years,
though, is the duration of the right in the bill after death.
I'm talking about the duration of a transfer even during life.
So, if you had, say, a 21-year-old artist who's granting a
transfer of rights in their image, likeness, or voice, there
should not be a possibility of licensing that for 50 years or
60 years during their life and not have any ability to
renegotiate that transfer. So, I think there should be a
shorter, perhaps 7-year, limitation on those transfers.
Senator Hirono. Makes sense.
Mr. Sheffner. Senator, one thing that we do like about the
bill is the First Amendment exemptions. We think they are most
of the way there to giving our members the clarity and
certainty they need. I think they can be improved a little bit,
and we have some specific, fairly technical changes that we
recommend. One thing that we would recommend changing is there
is currently essentially a no-preemption provision. We think it
should essentially be the opposite. I was just discussing with
Senator Tillis: this novel law with kind of some uncertainties
around the First Amendment limits. We think it would be
important not to preempt all existing State right-of-publicity
law but to preempt State regulation of digital replicas in
expressive works like movies, TV shows, songs, that are
protected by the First Amendment.
Mr. Davies. Thank you for the question. I think----
Senator Hirono. Do you mind, Mr. Chairman, if we----
Chair Coons. Yes.
Senator Hirono [continuing]. Just----
Chair Coons. I think this Is----
Senator Hirono [continuing]. Continue with the----
Chair Coons [continuing]. Instructive.
Senator Hirono [continuing]. Responses? Go ahead.
Mr. Davies. Yes. Building on some of the things said, I
think the efforts here to protect personhood is something we
very much encourage with the draft. The fact it's a discussion
draft--I think in terms of the key areas that we want to focus
on is where liability sits and that we would encourage it to be
focused on the creator and those that are first releasing the
content. We would prefer that it was based around right-of-
publicity laws, the existing body of law, rather than IP;
actual damages rather than statutory damages; and the
preemption message.
Senator Hirono. Professor.
Professor Ramsey. All right. Thank you for your question.
So, what I like most about the bill? I love the specific
exclusions from liability, even though there might be some
additional revisions that should be made, and the fact that
you're protecting personhood, although I'll note that State
right-of-publicity laws do sometimes apply to noncommercial
uses of a person's identity. The Zacchini case involved use of
his entire act in a news report, so that's not a commercial
use. And in the Comedy III case in California, the Supreme
Court case, the law was applied to an identical kind of
rendition of The Three Stooges in a lithograph, which is also
not commercial speech. So, there are some circumstances where
current laws do apply to noncommercial uses of a person's
identity, and also false endorsement laws can apply to
noncommercial uses, but they have to be used in connection with
goods and services, so they might be noncommercial use in goods
and services.
What should we change? Well, I think there's a tie. You
said pick one, but I have to pick two. So, first I think we
should have separate causes of action, as I mentioned before,
with distinct defenses. So, for example, a disclaimer might
make sense if you're targeting deceptive impersonation of
someone, right, because it dispels any confusion. But a
disclaimer doesn't make sense in a sexually explicit deepfake
that's been put out there without consent. You might have
different requirements with regard to commercial use, right? If
it's just a general, broad Federal right-of-publicity, you
might have a commercial use requirement, whereas if you're
talking about sexually explicit deepfakes, impersonation, it
should apply to both commercial and noncommercial speech.
And then my other part of the tie is the provisions with
regard to limits--or no limits on the scope of licensing. My
concern is that individuals without significant bargaining
power at the early stages of their career might sign a
contract, maybe a long contract, where it has a digital rights
provision in it, and sign away the right to their identity for
a lengthy period of time and use in any context.
And so I would like to see some way for Congress to
encourage or require, you know, those folks who are negotiating
these agreements to perhaps have a specific use authorization
for a certain movie, right, as opposed to use of your digital
identity in any context, right. Or instead of a lengthy period
of time--instead of 50, 60 years, I would say maybe 1 to 5
years. And I'm not an expert in the area of what's a good term,
but I think it's critical to make that shorter rather than
longer, because a lot of these people, even if they have
attorneys, they're just not going to have that kind of
bargaining power that the big studios, big music companies will
have.
Senator Hirono. Thank you, Mr. Chairman. I think that what
the Professor is suggesting, different causes of action--it's
very intriguing but complicated, so----
Chair Coons. Perhaps we'll explore----
Senator Hirono [continuing]. We'll think on it.
Chair Coons [continuing]. Second round.
Senator Hirono. Thank you very much.
Chair Coons. Senator Blackburn.
Senator Blackburn. Thank you, Mr. Chairman, and thank you
for your good work on the bill. We've spent months working on a
discussion draft and moving this forward, so I'm so pleased
that today we are to the hearing stage on this. Now, I
represent Tennessee, so it doesn't matter if you're on Beale
Street or if you're on Music Row or maybe you're working with
Naxos or one of the symphony distributors--you know, we
distribute more symphonic music out of Nashville, Tennessee
than anybody else in the world. We've got gospel, contemporary
Christian. We've got church music. We've got bluegrass. We've
got the Museum of African American Music. It is all right
there. And we are really so protective of our creators.
And in Tennessee we kind of have the good, the bad, the
ugly relationship when it comes to AI. All of our people in
manufacturing and logistics and health care--they're innovating
and going to town with it, but I'm deeply concerned about what
is happening to the creative community and working on NO FAKES
and making certain that there is a way for artists to protect
that name, image, likeness, their voice--that there is a way
for them to exercise their constitutional right to protect
their intellectual property and to benefit from that property.
That is going to be so important.
Mr. Kyncl, I want to come to you. I appreciate the comment
you made when we were visiting, preparing for the hearing. You
said, we got data wrong, data privacy wrong. We still haven't
done data privacy.
Mr. Kyncl. Yes.
Senator Blackburn. And we can't afford to get AI wrong. And
it's going to require that we take an action. And Tennessee
stepped up last month, and they passed the ELVIS Act. This is a
great piece of legislation. And, Mr. Chairman, what they did
was take much of what we've put in the discussion draft, and
they put it in place to protect our innovators and to give them
that State right of action. And not all States are following
suit on this, of course, and I think that what we've done is
kind of establish that baseline for a Federal action. So, I'd
like to hear from you, if you will, sir, about the need for a
Federal standard, a Federal preemption on that, for action.
Mr. Kyncl. Yes. Thank you for your efforts on the ELVIS
Act. It's truly groundbreaking. We are in a unique moment of
time, where we can still act, and we can get it right before it
gets out of hand. The genie is not yet out of the bottle, but
it will be soon. And, as you mentioned, Senator Blackburn, we
got it wrong on privacy. We waited too late. Don't get it wrong
on identity. It's simply far too important. And the speed at
which this will happen will be afforded by open-sourcing of
foundational AI models which are developed. And once that
happens, everything accelerates exponentially, and therefore
it's imperative that Congress acts this year. Thank you.
Senator Blackburn. You know, we've heard--some commentators
are talking about, well, you've got existing law, when it comes
to privacy or to personal property and intellectual property
protections, and so you can rest on that existing law, and that
is sufficient to go in and get a takedown order on some of
these AI fakes. Talk to me about why that is not sufficient.
Mr. Kyncl. I mean, today, if you think about privacy, how
many spam emails do you get every single day in your inbox?
Quite a lot. Your personal information is leaking everywhere,
whether it's being sold or whether it's just being taken. It's
just not safeguarded properly. When that happens with your face
and your voice, it's a whole new game. And for you--and this
will happen at a volume that is impossible for every single
person to try to, like, personally manage, which means it has
to be solved with technology. So, it is technology that will
unleash it, and it has to be technology that helps manage it,
which is why it's important for us to work with the technology
platforms to solve this, and we have to have a working bill and
working law that can be operationalized by all of us. But the
existing framework is simply too whack-a-mole and doesn't work.
Senator Blackburn. Let me ask you this. And, Mr. Chairman,
if I can get one more question in. Do you think that the
platforms should be held responsible for unauthorized AI fakes
that they're continuing to allow to be distributed?
Mr. Kyncl. I think we need to develop a set of conditions--
--
Senator Blackburn. Okay.
Mr. Kyncl [continuing]. That they should meet, and then if
they don't, then yes. But there has to be an opportunity for
them to collaborate and work together with all of us to make it
so. And that, I think, is the detail work that needs to happen.
But when we achieve that, then it will work, and there will be
good actors, and many of them are. So, I think it's through
that collaboration that we can wrestle this down.
Senator Blackburn. Thank you. Thank you, Mr. Chairman.
Chair Coons. Thank you, Senator Blackburn, and thank you
for your cooperation on moving forward this great bill. I have
a whole series of questions I want to ask about potential
tweaks, so I'm going to try and move relatively quickly, if I
might.
Mr. Sheffner, you testified that we have to include First
Amendment exceptions for uses and works that have public
interest or newsworthy value. Some people say that any work
involving a celebrity is newsworthy or in the public interest.
And that raises the challenge of how we define First Amendment
exceptions to ensure they don't just swallow up the rule and
permit all kinds of uses that the bill's, in fact, trying to
stop. So, I'd be interested in your views on how we narrow
that. And, Professor Ramsey, how would you craft the First
Amendment exceptions to make sure that they don't swallow up
the whole bill, particularly with regards to what is
newsworthy?
Mr. Sheffner. Sure. So, Senator Coons, we've talked to your
staff, which I have a great relationship with; I've talked to
other stakeholders and listened to the concerns that've been
raised about, well, maybe these exceptions are overbroad and
they could somehow swallow the right itself. We've listened,
and we've suggested tweaks to make sure that those types of
exceptions do not apply if the use of the digital replica is
deceptive. We don't support fraud. Fraud is not protected by
the First Amendment. It should not be allowed.
Chair Coons. Does----
Mr. Sheffner. One----
Chair Coons. Does it----
Mr. Sheffner. One--well, but one other thing I would just
say is that these types of statutory exemptions have been
routinely included in State right-of-publicity laws over the
last 25 years or so, since the late 90s. And one thing that
we've seen--or that we haven't seen is this type of abuses of
those exceptions. They have worked very well in separating out
the uses where you should need to get permission--again, to put
somebody's face on a billboard or on a lunchbox----
Chair Coons. Got it.
Mr. Sheffner [continuing]. Versus the biopics.
Chair Coons. Got it. Professor. Briefly.
Professor Ramsey. So, Christine Farley and I just recently
wrote a paper about how we can balance trademark and free
speech rights when someone uses a trademark in an informational
or expressive way, like a news report, entertainment, things
like that. And I think our proposal in that context might also
work here. So, as you mentioned, right, some of these kinds of
uses can actually be bad, right? Impersonation, et cetera.
So, one approach, in addition to listing out these
potential defenses, would be to say, if this is an
informational expressive use that is a false statement or false
representation--so you actually say this is a certain
celebrity, when it's not, or a certain teenaged girl, when it's
not--that would be actionable, still, even though, you know,
there's some argument that it's expressive, or if this use is
likely to mislead a reasonable person about the source of the
message or the speaker's identity. And so that way, you would
be able to at least have courts consider whether it's an
information or expressive use but also the safety valve that if
it's really causing harm because it's deceptive, then you can
still regulate it.
Chair Coons. Understood. Mr. Davies, today you raised
concerns the bill lacks a mechanism for showing or
demonstrating your members had knowledge. Should we incorporate
a notice-and-takedown structure? If so, should it be the DMCA's
notice-and-takedown provisions? Is there another mechanism
you'd urge us to consider for knowledge and for constructive--
--
Mr. Davies. Yes. Thank you for the question. In terms of
the current situation, our members are handling--they're the
leading streaming services, so they're handling the majority of
the music streaming consumption, and the processes are working
very well. I think the example that you've used and other--you
know, the Drake example, which is a common one--there's been no
challenge there in taking down the content expeditiously. So,
we don't see our members needing any additional burdens or
incentives here, but we do understand that the Committee is
keen to look at--if there is to be secondary liability, we
would very much seek that there be a safe harbor for an
effective takedown.
I think the DMCA notice-and-takedown process we don't see
as being a good process for here. It was designed for
copyright, and we obviously have opposition in terms of seeing
this as a different set of rights. That said, our members
absolutely can work with the Committee in terms of what we
would think would be an effective notice-and-takedown, and
building on some of the points the Professor's made there, in
terms of, it's really essential that we get specific
information on how to identify the offending content so that it
can be removed efficiently. We need information on the
notifier, in terms of why is the content offending, on what
basis, and also that information on the notified, so that if
there was an objection to the notification, that can take
place.
Chair Coons. Two more questions. I want to talk about
preemption briefly, Professor, if I might. Several witnesses
have described the existing State right-of- publicity laws as a
difficult-to-navigate patchwork. Should our bill broadly
preempt State laws or limit preemption to those State laws
governing uses of unauthorized digital replicas?
Professor Ramsey. So, I teach right-of-publicity law and
trademark law, intellectual property survey, and every year I
teach it, I think, we really need a Federal right-of-publicity
law, right. The State laws are so different. And if you go to--
Jennifer Rothman has this great blog which talks about all the
different laws. And then even within a State, the statutory
provisions have different rules than the common law provisions.
So----
Chair Coons. I take it your answer is yes?
Professor Ramsey. Well, yes.
[Laughter.]
Professor Ramsey. So, I'm just building it up. So, yes, we
need preemption, right. But the challenge, right, is, you know,
obviously, Congress--you're doing a great job trying to get
this right. And so you get it right, and then you preempt State
laws, and it simplifies everything: for litigants, for judges.
Instead of having to figure out which law's going to apply in a
particular--you know, there's right now forum shopping going
on, right. People are going to file suit in whatever State's
going to be best for their interests. And so if you have--yes.
So, yes, we need--preemption's a great idea.
Chair Coons. The discussion draft has a 70-year postmortem
provision. It's modeled after the Copyright Act. Postmortem
rights are important, but we understand 70 years is a long
time, especially for individuals who don't commercialize their
image, voice, or likeness. I'd be interested--jump ball.
Several of you; perhaps Mr. Sheffner first and then others.
Should postmortem terms be longer for individuals who
commercialize image, voice, and likeness? Should they be
limited? Should they be reviewed and re-extended every decade
or so? Like, how would you handle postmortem rights? The draft
has 70 years postmortem, and some of you have enthusiastically
supported that as part of your creative legacy. Others have
raised concerns. Mr. Sheffner, you kick us off, and we'll do
this one quickly.
Mr. Sheffner. Sure. So, we view this, again, through the
lens that this is a content-based regulation of speech. And as
Professor Ramsey said in her opening statement, a content-based
regulation of speech needs to be justified by a compelling
government interest that is narrowly tailored to serve that
interest. And what we have said is that, as for living
professional performers, use of a digital replica without their
consent impacts their ability to earn a living. You have a
compelling government interest in regulating there, and it
would be appropriate for Congress to regulate. Postmortem, that
job preservation justification goes away, and I have yet to
hear a compelling government interest in protecting digital
replicas once somebody is deceased. So, I think there's going
to be serious First Amendment problems with extending a right
that applies in expressive works postmortem.
Chair Coons. Any other witnesses think preserving the
legacy and property rights of an individual is worthy of some
protection? Professor Ramsey, and then Mr. Crabtree-Ireland.
Professor Ramsey. So, this is not going to shock you, but
I'm going to say it depends on the goal of the law, right? So,
if we're----
Chair Coons. You really do belong as a----
Professor Ramsey. Right.
Chair Coons [continuing]. Professor.
[Laughter.]
Professor Ramsey. So----
Chair Coons. Great.
Professor Ramsey. So, if we're talking about a law that's
regulating deceptive uses of someone's identity; if we're
talking about a law that's governing sexually explicit
deepfakes, it seems to me that it's fine to have a postmortem
right, like long--right, long term, possibly, maybe life plus
70. If we're talking, though, just about protection of a broad
Federal right-of-publicity, maybe not so much. And I haven't
written in this area, but I would recommend looking at the
works of people who have: Mark Bartholomew; Jennifer Rothman I
think is working on a paper, et cetera.
Chair Coons. All right. Mr. Crabtree-Ireland.
Mr. Crabtree-Ireland. Thank you. I mean, to me, it's
shocking that anyone would think that this right doesn't
deserve to be preserved and protected after death, I mean, for
all the reasons that twigs stated about how personal this is.
It's an economic right, it's a personal right, and it's
something that has real value. And so why that should somehow
dissipate upon death and make itself available to big corporate
interests like the ones represented by some folks here--that
doesn't make any sense.
I would argue that there shouldn't be a 70-year limitation
at all. This right should be perpetual, and the reason why this
right should be perpetual is that every one of us is unique.
There is no other twigs, and there never will be. There is no
other you, or you, or any of us. This is not the same thing as
copyright. It's not the same thing as--we're going to, you
know, use this to create more creativity on top of it, later.
This is about a person's legacy. This is about a person's right
to give this to their family and let their family take
advantage of the economic benefits they worked their whole life
to achieve. So, from my perspective, this is an intellectual
property right that deserves protection. It should absolutely
be protected after death, and I'm waiting to hear a good reason
why it shouldn't be, to be honest with you.
Chair Coons. In perpetuity; not at all. Mr. Kyncl, see if
you can help us bring this home, and then I'm----
Mr. Kyncl. I will make this----
Chair Coons [continuing]. Going to----
Mr. Kyncl [continuing]. Very brief for you. I agree with
Mr. Duncan Crabtree-Ireland 100 percent.
Chair Coons. Thank you all for that. Twigs, would you like
to make a comment on that? Forgive me.
Ms. Barnett. I was going to say that I've worked so hard,
throughout the whole of my career, and when I die, I would like
everything that I've created to go to my family and my estate,
that will have clear instructions of the way I want to preserve
my history and all of the art that I've created.
Chair Coons. Thank you. Senator Blumenthal.
Senator Blumenthal. Thank you very much, Mr. Chairman. I
got off a plane about 20 minutes ago, coming from Connecticut,
so I do apologize for missing the bulk of the hearing. As you
may have heard, we had no votes yesterday, so today was a
partial day off, and I had plans in Connecticut. So I am
grateful to all of you for being here, and we are very, very
hopeful that you're in good health and that you're going to
continue creating, and I'm a big fan of your work. So, thank
you for being here, particularly.
Thank you, Mr. Chairman, for having this hearing, which
focuses on a bill that you're going to introduce. I'd like to
be added at the appropriate time as a co-sponsor. I'm a strong
supporter, and I believe that there ought to be a Federal right
for people whose image and voice are used without their
consent, whether it is an actor or a songwriter or a singer or
an athlete. What is shared here is a right in one's own
likeness and creation as a person, an individual right. And I
think there ought to be a right to take legal action under that
right. A right without a remedy is unavailing, as we know from
our first year in law school, which for me was quite a few
years ago, but I've seen it repeated again and again in real
life as a prosecutor, as an advocate, as a litigator.
But I'd also like to focus on a complementary remedy, which
could be watermarking or identification, attribution, giving
credit, not just the deepfake and the right to recover as a
result of use of it without attribution or credit, so to speak,
without a watermarking, but also that kind of identification,
public crediting of a work. And I'm asking not only in the
abstract, but I head a different Subcommittee of the Judiciary
Committee. It's called Privacy, Technology, and the Law.
The Ranking Member of that Subcommittee and I, Senator Josh
Hawley of Missouri, have set forth a framework. It's the most
comprehensive bipartisan framework right now, and we should do
more adopting the kind of measure that Senator Coons and others
have proposed. But it would provide a requirement for
watermarking, as well as an entity to oversee licensing,
mandatory licensing, risk-based, of AI models and other
measures like transparency and so forth. It's a more
comprehensive approach.
But my question is really focused on watermarking. Maybe
you can tell us--let me ask all the witnesses--how watermarking
can complement rules requiring permission to use someone's
likeness or voice or creation in a deepfake or impersonation or
simply using it without permission.
Mr. Kyncl. I'll take it. So, thank you for all of your work
on this important issue. I think, you know, without attribution
achieved through watermarking, we won't be able to
operationalize what we're talking about here today. So, you're
focusing on absolutely the right issue. And I think the
important part in this is to determine the provenance of
content that's being displayed, the degrees of similarity to
its original--you know, to the original, and then it is up to
the rights holders, whether it's artists, music companies,
movie studios, et cetera, to then negotiate commercial
relationships with the platforms, separate and aside from the
laws and how it all works, using all of those mechanisms.
We've actually done this. When I was at YouTube, this is
precisely what we have done with user-generated content. We've
just done it in the copyright scheme, where it was the exact
content referenced. And so we built a whole framework around
that. This is merely that, on steroids, adopted for the AI age
with, you know, many more shades of grey and much more speed.
But it's really just upgrading that. But the framework exists.
It has been developed by companies like YouTube, which is best
in class in that. And therefore I'm hopeful that we can take it
further and apply that to AI as it relates to voice and degrees
of similarity, using watermarks to label content and carry the
provenance of it.
Senator Blumenthal. Thank you.
Ms. Barnett. I mean, I can only really talk from personal
experience, that in the last 6 months, I had 85 of my songs
leak online, which was basically the whole of my
experimentation for my next album. It was really scary, because
I thought that it's like having the whole of my note pad, I
guess, of all my ideas being put out to the whole world before
it was ready.
But on the flip side of that, I felt very secure because I
was able to call up my label and say, hey, this has happened,
and immediately they could go and take it down, and it just
disappeared. And now you can't find it. So, I think that
watermarking--it would protect artists, because then we'll have
a point of call to go to, to say, this has happened, and
immediately whatever's, you know, been leaked online or put
online can be taken down.
But one thing I will say is that the thing that's really
scary is once something is out in the world, we can't take it
back. So, if someone uses my likeness and says something
offensive or says something harmful, people might think that
that is me. And we've all seen in the news when someone does
something wrong, and the big story is, like, the front page,
but then you think, oh, no, they actually didn't do something
wrong. It was a mistake. And the rewrite of it is so small, and
I think that's the thing that I'm scared about--is even if
something does get out in the world that's not me, it's the
reputational damage that it will do and the financial and
cultural harm that won't be able to be amended after the fact.
Senator Blumenthal. Very good point. If the chairman would
give me a little more time, I'd be interested in the others to
answer. Thank you.
Mr. Crabtree-Ireland. Thank you. I agree with Mr. Kyncl on
the value of watermarking and other tools, as well--C2PA: the
coalition is working on that--but I also just want to caution,
especially in deepfakes, it was mentioned earlier the idea of
disclaimers solving problems there, or the ideas of
watermarking solving problems there. We also have to make sure
that tools that we use to protect against abuses of these
technologies are realistic.
And so expecting viewers of content online to read deeply
into captions to find disclaimers or things like that--that
doesn't really solve this problem. So, I hope as the Committee
considers what to do, it's not enticed into thinking that that
type of solution actually solves the problem. It needs to be
more front facing, so that the message that's delivered is
received by all those who view it.
Mr. Sheffner. Thank you for the question, Senator
Blumenthal. As Mr. Kyncl was talking about in the copyright
context, watermarking has proved useful in certain contexts. I
think he was referring to YouTube's content ID system, which
has been a great help in reducing the presence of pirated
material on that platform. I would just say, again, our
experience from copyright law, though--it's not a silver
bullet. It sometimes can help identify the original source of
pirated material, but just because it's out--just because it
has a watermark on it, it doesn't stop it from being further
disseminated, et cetera. So, there's really no silver bullets
in this context.
Mr. Davies. Thank you for the question. I'm going to build
on things that have already been said. I think Robert talked
about the partnerships between the services and the rights
holders. These are absolutely essential. This is where the
content comes from, for the services. So, we're very reliant on
the data, on the metadata that exists. I think it would be true
to say that data in the music industry already have significant
challenges, so--but these are challenges we work on together.
Senator Blumenthal. Thank you. Professor.
Professor Ramsey. So, I'll incorporate by reference
everything that's been said before but then also say that I
think someone using a digital replica to impersonate someone or
basically put out a sexually explicit deepfake--they're not
going to use this kind of technology, so it's not going to help
in certain circumstances.
Senator Blumenthal. Yes. And I meant this--I think I used
the word complementary. If not, I meant to say complementary. I
didn't mean it as a substitute. So, I take all these comments
as very helpful and valid. Thank you, Mr. Chairman.
Senator Tillis [presiding]. On behalf of the Chair, Senator
Klobufar--char.
Senator Klobuchar. Thank you very much. That was an AI
attempt. I know it was.
[Laughter.]
Senator Klobuchar. Kind of failed. Kind of close, not
quite.
Senator Tillis. I never heard----
Senator Klobuchar. Okay. Professor Ramsey, since you ended
there, I'll pick up where you were, about some of these, and
some of the other witnesses mentioned about this--deepfakes and
how some of these things--whether it's sexually explicit images
or whether it is the political robocalls or videos or ads. And
I wasn't going to start this way, but it makes sense here
because of what you just said.
To me, some of this we just have to get off there. They're
not going to be able to listen to a major candidate for
president for 3 minutes and then look and see a label. And I
think that, in other countries, that's what they've done.
That's why Senator Hawley and Senator Coons and Collins and a
number of other Senators have come together. We're marking up
this bill, along with a labeling bill, in the Rules Committee
on elections. Could you talk about why that kind of targeted
approach to some of these, like, hair-on-fire things is very
important, given the timing of all of this?
Professor Ramsey. Well, as you can expect, I love the fact
that you're working on these targeted laws, but again, one of
the things we need to do is protect ordinary people from
impersonation. Over Thanksgiving, someone called my dad when I
was standing right next to him. It sounded just like my
brother, and he said he was in jail and he needed money to get
out of jail. And my dad was not duped by this, but, you know,
some people have been, as the Senators have noted. So, I think
it's a great idea, but I think that, you know, we still need
the more broad act to deal with these kinds of issues for folks
that are not politicians, et cetera.
Senator Klobuchar. Exactly. And my State director's son is
in the Marines, and her husband got a call where it was an
impersonating. They'd scraped his voice. They didn't know where
he was stationed. So, we're going to see all of this deployed
against military families, as well, really, all these kinds of
scams. So, it's going to be--I see this, you know, heaven of
some of the great uses, especially in health care, of AI, but
then there's the hell part, and that should be our job to try
to put the guardrails in place, which is why I'm so honored to
be working with Senator Coons and Tillis and Blackburn on this
bill.
So, one of the things that interested me during the
testimony--you, Mr. Sheffner, and Mr. Crabtree-Ireland, you
kind of got to this, but both the NO FAKES Act and this
election bill include exemptions, exceptions for the use of
digital replicas, to ensure the bills do not chill speech
protected by the First Amendment. Could you talk a little bit
more as we look at how we can write these in a way--and I have
tried, with exceptions for satire in the elections bill, with
Senator Hawley--how we can do this to ensure that common-sense
safeguards do not chill protected speech and that this is
upheld in a court?
Mr. Sheffner. All right. So, Senator Klobuchar, I just want
to say, agreeing with Professor Ramsey, that I think your
approach of having specific legislation on pornographic
deepfakes; other legislation on election-related deepfakes is
really the right way to go. When you have a broad bill that
essentially says you need permission to use digital replicas
and then let courts kind of sort it all out, that's where you
get into trouble, and you have an overbroad bill that is going
to necessarily end up encompassing protected speech. Makes it
vulnerable to being struck down on overbreadth grounds.
So, these kinds of exceptions, I think, are specific to the
type of legislation. In the world of movies, our studios, the
studios that we represent at the MPA, make a lot of movies that
are based on or inspired by real people and events. Went
through this morning. The last 5 years of all the Best Picture
nominees over the last 5 years--approximately half were based
on or inspired by real people and events. Our studios want to
make sure that legislation like this doesn't interfere with
their ability to do that.
When you're talking about, say, nonconsensual pornographic
deepfakes, you don't need those exceptions for biopics and
satire and parody. That stuff is bad in almost every
circumstance you can think of. And I think this narrow,
targeted approach is really the right way to go.
Senator Klobuchar. Okay. So, Mr. Duncan Crabtree-Ireland--
you've got, like, the best long name in the world.
Mr. Crabtree-Ireland. Thank you.
[Laughter.]
Senator Klobuchar. Could you talk about balancing that
right of creators with the right of those whose voice or
likeness may be at risk? Sitting next to one of them, right
there, with----
Mr. Crabtree-Ireland. Absolutely.
Senator Klobuchar [continuing]. Twigs. And how do you
believe we should balance that?
Mr. Crabtree-Ireland. Absolutely. You know, I think we all
agree that obviously the First Amendment has to be protected
and that expressive speech is important. I think, you know, the
exceptions that are written into this discussion draft now are
not that far off, but I think it's important that they not be
expanded upon nor that they be broader than necessary, because
the fact is, we can't anticipate what this technology is going
to do tomorrow. We cannot anticipate every iteration of this.
And while there are certain specific uses or concerns that are
being addressed by legislation like the legislation you've
referenced, there is a broader need for protection. The example
I gave in my opening statement is one. Twigs has given examples
as they apply to her.
And so we do need to have that proper balance, and I am
concerned that we are only looking at one side of the First
Amendment consideration here. The other side of the First
Amendment consideration is the right that each of us has to our
own freedom of speech, to be able to communicate our ideas, to
associate ourselves with ideas that we want to associate with
and not be associated with ideas we disagree with. And that is
being really trampled on right now by this unfettered ability
of people, without a Federal right, to do things like the
deepfake I experienced, that she experienced, et cetera.
And so I do feel like the Committee is going to have to
work on, you know, defining these exceptions, making sure they
are no broader than necessary to keep the legislation viable
but also to make sure it doesn't swallow up the rule, like the
Chairman said. If we make them so broad that they swallow up
the rule, then all of this work will have been for nought, and
the reality is, today is not like 10 years ago. It's not like
30 years ago. This technology is fundamentally different, and
what it can do with all of our faces and voices calls out--it
screams out for a remedy that's actually effective.
Senator Klobuchar. And do you see--and maybe anyone, twigs,
any of you--Mr. Kyncl--and again, this need for a national
standard, just because Senator Blackburn's worked with us on
this bill and is going to be a co-sponsor, and they just did
the ELVIS Act. Of course, in Minnesota we have the DYLAN Act
and the PRINCE Act. No, I just made that up.
[Laughter.]
Senator Klobuchar. But we do have people, as you know, who
are fiercely, fiercely independent and protective of their
incredible music in our State, but we have a common law in
Minnesota that's helpful. There's like, this State, this State.
Talk about--a few of you, if you want to--just this need to
have this national standard and why it's so important.
Mr. Kyncl. Maybe if I can----
Senator Klobuchar. Okay. Mr.----
Mr. Kyncl [continuing]. Chime in.
Senator Klobuchar [continuing]. Kyncl.
Mr. Kyncl. I just want to comment on some of the things
from before, which is, as someone who grew up without the First
Amendment, I value it probably more than those who have,
because I do not take it for granted at all. And it seems like
it's well and alive in America, because half of the movies who
were nominated for Oscars, you know, were based on, you know,
existing folks. So, saying that any, you know, AI regulation
that is respectful of the existing First Amendment is not
reducing it; it's keeping it as it is, and it's alive and well.
So, I do think that we need to stay within the limits of First
Amendment and not go beyond.
As to national regulation, we work with global platforms.
We're talking about global platforms. Not even national. We're
talking about global platforms. Doing anything State by State
is a very cumbersome process. Twigs's content getting on a
platform unauthorized--if we have to fight that on a State-by-
State, it's untenable. It just doesn't work.
Senator Klobuchar. Very good. Mr. Davies--he'll be my last
one, and then we'll--go ahead.
Mr. Davies. Thank you. I just need to reinforce what
Robert's just said. You know, absolutely right. You know, music
streaming is global. The success of this is having access to
twigs's music from the UK or from Tennessee or wherever. So,
it's high volume, and anything that adds complexity on a State-
by-State level is anathema to this industry. So, we're very
strongly in favor of preemption.
Senator Klobuchar. Very good. Just the last thing, kind of
along those lines, is--don't laugh. I heard you, Coons. Just--
it'll be very fast, and you can put it in writing--of you, Mr.
Davies: In January we heard testimony that generative AI has
been used to create unauthorized digital replicas of news
anchors making comments, and we have a number of things going
on in the journalism area.
I have a vested interest. My dad was a journalist for the
Minneapolis Star Tribune, but also, Senator Kennedy and I have
the bill to push for negotiation of the content and to get them
reimbursed, mainly from Google and Facebook, for the use of
this content, something that's going on in Australia and
Canada, and I will not go on. But what steps can streaming
services take to ensure that unauthorized digital replicas of
journalists are not posted on the streaming platform?
Mr. Davies. Senator, if I could follow up with you after?
I'm----
Senator Klobuchar. Okay.
Mr. Davies [continuing]. Not briefed on that.
Senator Klobuchar. Okay. Excellent. Thank you.
Chair Coons [presiding]. Thank you, Senator Klobuchar. Back
to Senator Tillis for his second----
Senator Klobuchar. She had----
Chair Coons [continuing]. Round.
Senator Klobuchar [continuing]. One thing.
Chair Coons. Twigs, if you'd like to----
Senator Tillis. Go ahead, twigs.
Senator Klobuchar. Yes.
Ms. Barnett. Oh, thank you. I'd actually like to go back to
Mr. Sheffner's point about the desire to make very big and
financially successful films about artists without consent. I
think the problem is if you're able to use an artist's voice
and likeness without consent about their life story, you're
giving the impression that it's, I guess, the equivalent of an
autobiography rather than a biography, you know. And that's the
confusion, if you're able to use my voice and my exact face,
you're saying, this is what happened from my point of view, and
it's not. It's what happened from a team of writers in
Hollywood that want to overdramatize things and maybe make it
more tragic or, you know, more fantastical.
And I think that's what makes me really nervous and feel
uncomfortable and very vulnerable. I don't think it's fair
that, even after an artist is deceased that somebody would be
able to make a film about their life, using them. You know, we
can watch a film about a star from the past, and if it's an
actor, we know to take it with a pinch of salt. If it is the
person themself, then it just feels too unclear and not fair
and actually not in--what am I trying to say? Not the best
intention of the artist's legacy.
Senator Klobuchar. Thank you.
Chair Coons. Thank you. Senator Tillis.
Senator Tillis. Thank you, Mr. Chair. I'm going to be
brief. I did have a question for you, Mr. Crabtree-Ireland. The
current draft legislation--individuals only have a right to
license out their digital likenesses if they hire an attorney
or they're a member of a labor organization. We've gotten some
feedback, I think, that your organization in particular--that
this is a giveaway. Can you give me other examples in law, give
away or really giving--vectoring everybody either into legal
counsel or to your union? Can you give me examples, other areas
in law where this is the case, where you have to engage an
attorney or a labor interest to move forward?
Mr. Crabtree-Ireland. Sure. And I guess I would just say I
don't think it's just our union. It would be, you know, any
collective bargaining representative. But there are a number of
examples that occur in labor law, labor and employment law,
where there are defined worker protections that then can be
deviated from through a collective bargaining arrangement but
not through individual contracts. In this case, the proposal, I
think, is a little broader, a little more open because of the
option of securing individual representation by an attorney as
an alternative. That's not normally present in those kinds of
statutes. But I'm sure I could provide--I can't give you a
laundry list----
Senator Tillis. Yes, if you could----
Mr. Crabtree-Ireland [continuing]. Right now, but I could
provide some----
Senator Tillis [continuing]. For the record. We're going to
be submitting questions for the record, for all of your--
provide an opportunity for additional information.
Senator Tillis. Mr. Chair, I just think it's remarkable, if
you take a look at the attendance in the audience and the
engagement from the Members, you're hard pressed to see--I
mean, on certain subjects, but on technical subjects like this,
to have Members go twice their allotted time demonstrates the
interest. So, twigs, I'm going to end my questions with you. I
do believe that Congress needs to act, but you need to
understand that this is--it's tough to get virtually anything
done, even what appears to be common sense, for the reasons
that we've talked about. We're going to have constitutional
questions we have to address. We have to get to a number of
matters.
And hopefully we do get it done this year, but in your
opening statement you were emotional or appeared to be
emotional on one or two points, and I'm just trying--I think
that people need to understand--I think--excuse me--one of the
reasons you maybe got emotional is because this is an
existential threat to creators. And I'm trying to figure out
how we educate people on the difference between an original
creation from a human being and something that was either
created or augmented from a machine. And this is more of a
societal thing that we have to sort out. At what point is
society just prepared to say, boy, this sounds as good. I know
it comes from a machine.
You mentioned something about the investment of your fans,
that they've made in you. How do you invest in a relationship
with a machine? I mean, we're at an interesting point in time
in history, where we could have billions of people think the
inauthentic creation of a machine is somehow as good as the
hard work of a human being. So, I wonder at what point--and
when we lose all the creators--this is a philosophical
question--at what point can those machines never possibly match
the creative genius of an individual?
Ms. Barnett. Thank you.
Senator Tillis. That's okay. Red means on, which makes no
sense to me, but----
[Laughter.]
Ms. Barnett. I think that there's two things here. I feel
incredibly lucky to have spent the whole of my teenage years
without a smartphone. So, I straddle a generation where I
memorized all my friends' numbers. I would walk to my friend's
house. If we said we're going to meet at 1:00 o'clock, I just
would have to be there. You know, there was no texting and
saying that I was going to be late. I loved my brain, back
then. I loved how simple it was. I loved what truth was, back
then. I loved that I was able to think for myself.
Even where we're at with the internet now, it's so
confusing. You know, even if you just want to find a simple
news story, we can't. Even if you want to find the truth about,
you know, whether a food even is good for you or bad, we can't.
You know. It's just a stream of nonsense. I look at a lot of my
friends that have children that are teenagers, and their mental
health is really struggling. We're looking at young people that
have anxiety, that have depression because they're overwhelmed
with information and lack of truth and lack of stability.
And the thing that scares me is the--my fans look to me for
a north star, a message, a sense of being. My work is something
that they can find themselves in. And if you change the
narrative of my work, we're just messing with their brains. You
know, like, the solid essence of my work that I've spent 10
years developing--if someone can just take it and make up
something completely different, I'd feel so bad, because I'm
harming people then, and there would be nothing that I can do
about it.
I think the way that we can prevent this from happening is
putting the power in the hands of the artists and also putting
the power in the hands of people that are there to protect the
artists, whether that's third parties like record labels or
agents or lawyers. You know, that's up to the artist to
understand and to, you know, sign a contract if we want to. You
know.
But I think that the way that I've been experimenting with
deepfake is going to help my fans. It's going to help them
understand the nuance of my language across all parts of the
world. Like, the way that I want to use it is not harmful
because I think inherently artists just want to express their
emotions, be there for people, and say things that you can't
say for yourself. So, if you're putting words in our mouths,
it's going to be devastating.
Senator Tillis. Well, I also agree. I'm very glad there
weren't cell phones back when I was a young person, but----
[Laughter.]
Senator Tillis [continuing]. Maybe for other reasons--and
that Polaroids fade, but----
[Laughter.]
Senator Tillis. But, no, I do think that--you know, I'm
glad that we're taking up this bill. I do feel strongly that we
should do everything we can to try and move it in this
Congress. If not, then we just have to lean into it and get it
done in the near future. But when we have these discussions, it
points to all the other societal challenge, challenges of
creators that we need to get right.
This technology--I love it; I interact with generative AI
for about an hour every day, as part of my own study of it--and
study that began back in the 1980's into artificial
intelligence, for me personally--but we've got a lot of work to
do, and Congress has a role to play, but we've got to be very,
very careful not to overstep, not to trample the rights of
others. And we're going to need your help and your continued
engagement, to get it right. So, thank you all for being here
today.
Chair Coons. Senator Tillis, thank you. Thank you for,
again, being a great partner. I have even more questions, but
we have come to the end of our time. And you and Senator
Blackburn have been terrific to work with. I am grateful to all
of our witnesses for the way that you've brought your skills,
your value, your background, your creativity, your voice to
this hearing today. And we've engaged in a lot of different
challenging questions about how we could refine this, how we
could narrow it.
There've been a lot of members who've participated. For
those who did not participate or those who still have other
questions, the record will be open for questions for the record
for the witnesses. They are due 1 week from today by 5 p.m. on
May 7--although, twigs, in your case, 2 weeks before we wrap
this up with cellophane and move forward.
Chair Coons. If I could, today's hearing was important to
show that when we regulate the use of AI, we have to balance
individual privacy rights and First Amendment rights in a way
that doesn't stifle creativity and innovation with these
rapidly developing AI tools. It reinforces what we've heard
today, why we need a clear policy to protect image, voice, and
likeness of all individuals from unauthorized AI replicas. And
the feedback we heard and that our staff has received over the
last 6 months is critical.
I look forward to working with my colleagues and co-
sponsors and the witnesses and the others who attended today,
to refine this in the next week or two and get to the point
where we can introduce it next month, so we move from
discussion draft to reality. I think we need to seize the
moment and move forward. Thank you for your partnership. Thank
you for your testimony. And with that, this hearing is
adjourned.
[Whereupon, at 4:28 p.m., the hearing was adjourned.]
[Additional material submitted for the record follows.]
A P P E N D I X
Submitted by Chair Coons:
ESA--Entertainment Software Association--NO FAKES Act--Protecting
Americans from Unauthorized Digital Replicas.................. 142
IFTA--Independent Film & Television Alliance--Proposed NO FAKES
Act........................................................... 145
NAVA--National Association of Voice Actors--The NO FAKES Act..... 148
Submitted by Senator Tillis:
ESA--Entertainment Software Association--NO FAKES Act--Protecting
Americans from Unauthorized Digital Replicas.................. 142
IFTA--Independent Film & Television Alliance--Proposed NO FAKES
Act........................................................... 145
NO FAKES Act--Objection Letter................................... 150
[GRAPHIC(S) NOT AVAILABLE IN TIFF FORMAT]
[all]