[House Hearing, 118 Congress]
[From the U.S. Government Publishing Office]
OVERSIGHT OF THE U.S. COPYRIGHT OFFICE
=======================================================================
HEARING
BEFORE THE
SUBCOMMITTEE ON COURTS, INTELLECTUAL
PROPERTY, AND THE INTERNET
OF THE
COMMITTEE ON THE JUDICIARY
U.S. HOUSE OF REPRESENTATIVES
ONE HUNDRED EIGHTEENTH CONGRESS
FIRST SESSION
__________
WEDNESDAY, SEPTEMBER 27, 2023
__________
Serial No. 118-47
__________
Printed for the use of the Committee on the Judiciary
[GRAPHIC NOT AVAILABLE IN TIFF FORMAT]
Available via: http://judiciary.house.gov
__________
U.S. GOVERNMENT PUBLISHING OFFICE
53-722 WASHINGTON : 2023
-----------------------------------------------------------------------------------
COMMITTEE ON THE JUDICIARY
JIM JORDAN, Ohio, Chair
DARRELL ISSA, California JERROLD NADLER, New York, Ranking
KEN BUCK, Colorado Member
MATT GAETZ, Florida ZOE LOFGREN, California
MIKE JOHNSON, Louisiana SHEILA JACKSON LEE, Texas
ANDY BIGGS, Arizona STEVE COHEN, Tennessee
TOM McCLINTOCK, California HENRY C. ``HANK'' JOHNSON, Jr.,
TOM TIFFANY, Wisconsin Georgia
THOMAS MASSIE, Kentucky ADAM SCHIFF, California
CHIP ROY, Texas ERIC SWALWELL, California
DAN BISHOP, North Carolina TED LIEU, California
VICTORIA SPARTZ, Indiana PRAMILA JAYAPAL, Washington
SCOTT FITZGERALD, Wisconsin J. LUIS CORREA, California
CLIFF BENTZ, Oregon MARY GAY SCANLON, Pennsylvania
BEN CLINE, Virginia JOE NEGUSE, Colorado
LANCE GOODEN, Texas LUCY McBATH, Georgia
JEFF VAN DREW, New Jersey MADELEINE DEAN, Pennsylvania
TROY NEHLS, Texas VERONICA ESCOBAR, Texas
BARRY MOORE, Alabama DEBORAH ROSS, North Carolina
KEVIN KILEY, California CORI BUSH, Missouri
HARRIET HAGEMAN, Wyoming GLENN IVEY, Maryland
NATHANIEL MORAN, Texas BECCA BALINT, Vermont
LAUREL LEE, Florida
WESLEY HUNT, Texas
RUSSELL FRY, South Carolina
------
SUBCOMMITTEE ON COURTS, INTELLECTUAL PROPERTY, AND
THE INTERNET
DARRELL ISSA, California, Chair
THOMAS MASSIE, Kentucky HENRY C. ``HANK'' JOHNSON, Jr.,
SCOTT FITZGERALD, Wisconsin Georgia, Ranking Member
CLIFF BENTZ, Oregon TED LIEU, California
BEN CLINE, Virginia JOE NEGUSE, Colorado
LANCE GOODEN, Texas DEBORAH ROSS, North Carolina
KEVIN KILEY, California ADAM SCHIFF, California
NATHANIEL MORAN, Texas ZOE LOFGREN, California
LAUREL LEE, Florida MADELEINE DEAN, Pennsylvania
RUSSELL FRY, South Carolina GLENN IVEY, Maryland
CHRISTOPHER HIXON, Majority Staff Director
AMY RUTKIN, Minority Staff Director & Chief of Staff
C O N T E N T S
----------
Wednesday, September 27, 2023
Page
OPENING STATEMENTS
The Honorable Darrell Issa, Chair of the Subcommittee on Courts,
Intellectual Property, and the Internet from the State of
California..................................................... 1
The Honorable Henry C. ``Hank'' Johnson, Ranking Member of the
Subcommittee on Courts, Intellectual Property, and the Internet
from the State of Georgia...................................... 2
The Honorable Jerrold Nadler, Ranking Member of the Committee on
the Judiciary from the State of New York....................... 3
WITNESS
The Hon. Shira Perlmutter, Register of Copyrights & Director,
U.S. Copyright Office, Library of Congress
Oral Testimony................................................. 5
Prepared Testimony............................................. 8
LETTERS, STATEMENTS, ETC. SUBMITTED FOR THE HEARING
All materials submitted by the Subcommittee on Courts,
Intellectual Property, and the Internet, for the record are
listed below................................................... 42
QUESTIONS AND RESPONSES FOR THE RECORD
Questions to Register Shira Perlmutter, Register of Copyrights &
Director, U.S. Copyright Office, Library of Congress, submitted
by the Honorable Darrell Issa, Chair of the Subcommittee on
Courts, Intellectual Property, and the Internet from the State
of California, for the record
Response from Register Shira Perlmutter, Register of Copyrights
& Director, U.S. Copyright Office, Library of Congress
OVERSIGHT OF THE U.S. COPYRIGHT OFFICE
----------
Wednesday, September 27, 2023
House of Representatives
Subcommittee on Courts, Intellectual Property, and
the Internet
Committee on the Judiciary
Washington, DC
The Committee met, pursuant to notice, at 10:04 a.m., in
Room 2141, Rayburn House Office Building, the Hon. Darrell Issa
[Chair of the Subcommittee] presiding.
Members present: Representatives Issa, Jordan, Fitzgerald,
Cline, Kiley, Moran, Lee, Fry, Johnson of Georgia, Lieu, Ross,
Schiff, Lofgren, Dean, and Ivey.
Mr. Issa. [Presiding.] Good morning. The Subcommittee will
come to order.
Without objection, the Chair is authorized to declare a
recess at any time.
We want to welcome everyone here today, and particularly
note that there are conferences going on, and that a number of
members have been delayed, but will attend. So, please, it is
not you; it is the busyness of this week that causes a little
bit of a delay for some people being here.
I will now recognize myself for the joy of an opening
statement, which I will abbreviate.
As we all know, the Copyright Office plays a critical role
in our economy. Strong copyright protection, the incentives
that our creators, our Constitutional creators, bestowed, if
you will, not for the benefit of the creators, producers, or
innovators, but for the benefit of our society. The
Constitution itself confers a responsibility on Congress to
promote the progress of science and useful arts by securing for
limited times for authors and inventors the exclusive right to
their respective writings and discoveries.
We take that role seriously, and for more than 200 years,
have continued to debate what promotes. Most laws and most
things conferred in the Constitution are fairly static. They go
back their original intent. When the intent is, in fact, a
mandate to promote, we must from time to time review whether or
not that promotion is fair and appropriate. Too much, in fact,
we deny the public the opportunity to also share in those
creations. Too little, and those creations do not happen.
As the Copyright Office carries out a wide range of
responsibilities, it, too, must change with the time. When the
Copyright Office made the decision that, in fact, full AI-
produced had no right to a copyright, they did us all an
appropriate favor, because, in fact, we know that pushing a
button and letting that run for days, weeks, months, or years
is not, in fact, continuous innovation.
In fact, today, one of our discussions will be on the
intersection of artificial intelligence and copyrights, because
we do not want to limit AI from participating in the further
promotion of intellectual property that is useful in not only
copyright, but also in patent. We must make sure that this is,
in fact, an incentive, and not simply a reward for the first to
turn on a machine.
We recently held hearings on music modernization,
mechanical licensing, with the Copyright Office, and we look
forward to the potential renewal in 2025.
We also examined issues of the right to repair, including
examining the Copyright Office's Section 1201--hotly debated,
but necessary.
At each turn, we have been reminded that the importance of
the Copyright Office in all these areas cannot be overstated,
which is why I am pleased that the Register of Copyrights
accepted our invitation to appear at this hearing, so that we
can go straight to the source and explore the issues she faces,
in fact, whether there are appropriate intersections for us to
join in legislative changes.
I want to thank the Register, and I want to appreciate her
time.
So, with that, we will have an opening statement from Mr.
Johnson, and then, go right to our witness.
With that, I recognize the Ranking Member, Mr. Johnson of
Georgia.
Mr. Johnson of Georgia. Thank you, Mr. Chair, for hosting
and holding this hearing.
Thank you, Register Perlmutter, for your appearance today.
What does it mean for human beings to create? In a very
real way, this question is the subject of our hearing today.
This is not just a philosophical or academic query; it has
real-world implications for countless American families and
American businesses, and as such, our answer to this question
must be reflected in our laws.
For well over 100 years, our Copyright Office has protected
Americans' intellectual property, and by doing so, has
protected American creativity and innovation. That is no easy
task.
As technology evolves, intellectual property evolves with
it. Artists, entertainers, writers, musicians, coders, and
creators adapt to a changing world, and unfortunately, so do
those who seek to profit by stealing intellectual property or
by undermining copyright protections that are a cornerstone of
our free enterprise system and a driver of both art and
commerce.
This Committee has deep experience in examining new
technologies and determining how our laws must change to
acknowledge the role new methods of innovation play in the
world. Internet connectivity changed the way writing and art is
created and shared. So, in 1998, Congress passed the Digital
Millennium Copyright Act to change the way we protect creators.
When music streaming began, it disrupted our thriving music
community of artists, songwriters, publishers, and record
labels. Instead of letting it flounder, the Music Modernization
Act created new ways to ensure that songwriters are paid for
their work.
For creators, for the Copyright Office, and for this
Committee, the rise of artificial intelligence, or AI, might
pose the greatest challenge yet. AI models, trained on
copyrighted data, pose serious intellectual property questions
regarding licensing and credit for final products.
I'm glad the Copyright Office addressed the question of
whether AI--or excuse me. I'm glad that the Copyright Office
addressed the question of whether AI in their creations can
claim copyright protections and found that artists can only
seek copyright for their human contribution.
We all, however, know that this is just the beginning. The
European Union is considering finalizing legislation to
regulate the use of AI, and it includes regulations governing
so-called trustworthy or ethical AI. Other individual Nations
have also begun to take steps to say how they want AI to impact
their citizens' work, play--or how their citizens work, play,
and create.
The United States, the leader in AI development, stands
alone in its silence thus far. We cannot afford to ignore the
very real challenges AI presents, even as we enjoy its very
real benefits. Missing the moment would set our Nation back and
it would harm the artists who call America home.
I'm proud to represent the State of Georgia, also known as
the ``Hollywood of the South,'' where actors, screenwriters,
directors, and others live their dream. Georgia is also home to
a thriving music industry, where artists, songwriters, and
record labels live, work, and play.
Their work and the work of the ecosystem of creation that
encircles television and filmmaking, as well as music, from the
crew to the studios boosts the GDP in Georgia and keeps our
State strong.
I was glad to see the writers and the studios come to the
table and find common ground on many issues, including AI. It
is time for Congress to do the same.
Just before his death, Robert Kennedy told a crowd that,
You cannot measure a Nation just by its wealth. The Gross
National Product, measures neither our wit nor our courage;
neither our wisdom . . . nor our devotion to our country; it
measures everything, in short, except that which makes life
worthwhile.
The Copyright Office protects many intangibles that make
life worthwhile, and I sincerely hope that Congress has the
courage to protect creators as new technologies disrupt their
fields.
I would like to thank Chair Issa for calling this hearing
once again.
Once again, I would like to thank Director or Register
Perlmutter for being here. I look forward to hearing from you
about your agency's work.
Thank you.
Mr. Issa. I thank the gentleman.
I am now pleased to recognize the Ranking Member of the
Full Committee, Mr. Nadler, for his opening statement.
Mr. Nadler. Thank you, Mr. Chair.
Mr. Chair, I would like to begin by thanking Ms.
Perlmutter, the Register of Copyrights, for being here today.
The Copyright Office bears responsibility for registering
copyrights, examining copyright claims, and administering
statutory licenses--not to mention advising Congress on new and
emerging issues of copyright policy, among other
responsibilities.
As the Committee with jurisdiction over the matters related
to the American judicial system, including intellectual
property law, we have the responsibility to provide oversight
of this critical office.
Professional and amateur artists, authors, and coders rely
on copyright protection to put their creations out into the
world. These ownership rights promise that, with hard work and
a lot of luck, it is possible to make a living from one's
artistic and intellectual abilities.
Our intellectual property laws foster creativity in the
arts, as well as productivity and innovation. As such, the
Copyright Office serves a dual role in American culture; it is
protective of both the future and the past.
While giving our hits the space to age gracefully into
classics and for our visual artists the time to gain
appreciation, the Copyright Office must also wrestle with
integrating the newest artistic mediums and addressing
philosophical questions about the use of technology in human
creations.
What we determine holds value, and how we choose to protect
it, demonstrates to the world what we believe is important in
this country. It is often the men and women in the Copyright
Office who make those important determinations.
Today, much of our conversation will focus on the future.
Artificial intelligence, or AI, has changed and will continue
to change the way Americans create. We cannot escape the novel
difficult questions posed by the integration of AI models into
our work.
How should American creators interact with generative AI
tools? What regulations will guide our innovations as a Nation?
I was glad to see the Copyright Office address these questions
this year in its March guidance and August Notice of Inquiry on
the impact of generative AI. I look forward to hearing more
about the results of the Copyright Office's examination of how
AI will impact artists and intellectual property holders.
The Music Modernization Act, which this Committee led
efforts to enact in 2018, is an example of what can happen when
an entire industry agrees to work together to protect creators.
When streaming services arrived on the scene roughly a decade
earlier, they disrupted the music industry's normal way of
doing business. America's teenagers stopped waiting outside
stores for the latest star's album to drop. Instead, the
listening public moved online, where unlimited songs were
available at their fingertips, and unfortunately, where there
was no structure to ensure that songwriters could get paid.
By creating the Music Licensing Corporation, or MLC, to
ensure that songwriters receive the royalties they are due, we
in Congress made clear that, no matter how the music industry
evolves--and evolve it will--the people who write the songs
deserve to be paid for their work.
I'm looking forward to hearing from the Copyright Office,
as the entity responsible for the administration of the MLC, on
how it believes the MLC is accomplishing its goals.
Some protections simply never existed in the first place.
The United States is the only democratic Nation in the world
that does not pay its performing artists when their songs are
played on terrestrial radio. There is no reason not to pay our
creators for their work.
That is why I am proud to join Chair Issa in leading the
American Music Fairness Act, which would require broadcasters
to pay artists when their songs are played on the radio. Our
actions should match our values, and paying artists when their
songs are broadcast on AM and FM radio would take a vital step
toward that ideal.
The Copyright Office's jurisdiction is broad. Today will be
an opportunity to expand beyond AI and music. I am also looking
forward to hearing from Ms. Perlmutter about the status of the
implementation of the CASE Act; how the Copyright Office is
modernizing its outdated IT systems to better serve artists,
and what changes, if any, need to be made to Section 1201 of
the DMCA.
Thank you, Mr. Chair, for holding this hearing, and I yield
back the remainder of my time.
Mr. Issa. I thank the gentleman.
I now recognize--oh, sorry. Without objection, all other
opening statements will be included in the record.
It is now my honor to introduce our sole and important
witness. Ms. Perlmutter is the Register of Copyrights and
Director of the Copyright Office. She was appointed in October
2020. Prior to her appointment, she served as the Chief Policy
Officer and Director of International Affairs for the United
States Patent and Trademark Office.
We welcome our witness today and thank her for appearing.
I would ask that you please rise to take the oath.
Do you solemnly swear or affirm, under penalty of perjury,
that the testimony you are about to give will be the truth and
correct to the best of your knowledge, information, and
beliefs, so help you God?
Ms. Perlmutter. I do.
Mr. Issa. Thank you. Please be seated.
Let the record indicate the witness answered in the
affirmative.
Since you are the sole witness and the reason we are here
today, we won't hold you strictly to the five-minutes, but the
sooner we get to questions, the happier people here on this
other part of this very cold room will be.
I'm noting that because we asked to have it warmed up and
we will do our best.
Thank you, and you are recognized for your statement.
We will turn the mic on. I know it is a little cold this
morning, too.
STATEMENT OF THE HON. SHIRA PERLMUTTER
Ms. Perlmutter. There we go.
Mr. Issa. OK, and then, get it a little closer, and we will
all be happy. Thank you.
Ms. Perlmutter. Good morning, Chair Issa, Ranking Member
Johnson, and Members of the Subcommittee.
Thank you for the opportunity to update you today on the
recent accomplishments and current projects of the Copyright
Office.
The past year has been very productive on multiple fronts.
We have substantially improved processing times, while making
significant progress on IT modernization.
We've launched work on artificial intelligence; marked a
full year of operations of our new small claims tribunal;
produced a number of policy studies; developed an economic
research agenda, and engaged in rulemakings under the Music
Modernization Act, as well as commencing the next Section 1201
rulemaking.
The Office's law and policy activities have been wide-
ranging. Most notably, we have moved quickly to address the
copyright implications of artificial intelligence. Early this
year, we announced a broad AI initiative, and in March, issued
guidance on how to apply to register works that incorporate AI-
generated content.
That guidance reaffirmed our longstanding position that
human authorship is required for copyright protection--a
position that was recently upheld by the District Court for the
District of Columbia.
Over the past six months, we've held a series of public
listening sessions and webinars and have met with a diverse
range of interested parties.
We, then, published a Notice of Inquiry at the end of
August seeking public comment on a full range of copyright-
related issues, including the legal treatment of the ingestion
of copyrighted works for machine learning; the copyright-
ability of the output, and the imitation of the likeness and
style of human creators.
As of the first anniversary of the small claims tribunal,
the Copyright Claims Board that was established by the CASE
Act, nearly 500 claims have been filed--with about 10 percent
of them so far in active proceedings. We've seen a steady
influx involving a wide range of types of works with strong
participation by individuals appearing pro se.
Over the past year, the Office has responded to
congressional inquiries on a number of issues, including
studies on deferred examination of registration applications;
on electronic deposits and the best edition requirement; and on
standard technical measures, as defined in the Digital
Millennium Copyright Act.
We are currently completing a study on non-fungible tokens
and intellectual property jointly with the Patent and Trademark
Office.
In our role administering the Copyright Act, the Office
registered over 484,000 claims to copyright for millions of
works in fiscal 2022.
Processing times are at a historic low. The average for all
copyright claims now stands 2.1 months. For fully electronic
claims that don't require correspondence, the average is just
over 1 month.
We also recorded more than 14,000 documents containing
titles of over a million works. A major milestone was reached
last year with the opening of our online recordation pilot to
the public. Almost 80 percent of all basic recordation
documents are now submitted online, and processing times are
measured in weeks rather than months.
We've completed the consolidation of deposit materials from
several facilities into a single modern warehouse to enable
faster location services that are tracking and improved
security.
We're also implementing a new multichannel contact center
to enhance communications with the public.
Modernization remains a top priority. The planned
Enterprise Copyright System, or ECS, will update and connect
all our services, making them more efficient and easier to use.
The first public releases were the online recordation system
and a pilot of our new Copyright Public Records System.
Development of a new and improved registration system is
now well underway, and we've made considerable progress in
digitizing pre-1978 records and making them available online.
As the ECS becomes fully operational, we will focus on
continuous development with regular maintenance and updating to
avoid repeated overhauls of legacy systems.
Finally, we've expanded our outreach through a greater
range of educational materials and events offered to more
audiences to further our goal of copyright for all.
So, let me close by thanking the Subcommittee for your
support of the Office's work to foster creativity and promote a
thriving copyright system.
[The prepared statement of the Hon. Perlmutter follows:]
[GRAPHICS NOT AVAILABLE IN TIFF FORMAT]
Mr. Issa. Thank you.
I'm going to forego my questions initially and go to Mr.
Fitzgerald.
Mr. Fitzgerald. Thank you, Mr. Chair.
Thank you for being here today.
September 5th, the Copyright Office recently published an
interpretative rule in the Federal Register relating to when
late fees apply under the Music Modernization Act, instead of a
proposed rule that would have been subject to public notice and
comment. Given the potential commercial impact and the
disruption of settled industry practices by this rule, can you
speak to why the Associate Register of Copyrights, Suzy Wilson,
is listed as the signatory on the rulemaking rather than
yourself?
Ms. Perlmutter. Suzy Wilson, as you said, ``Associate
Register and also General Counsel.'' It is not unusual for the
General Counsel to sign regulations that we issue in the
Federal Register. It had no other substantive significance
beyond that.
Mr. Fitzgerald. OK. Very good.
Do you or your office intend to issue further
interpretative rules that circumvent public comment periods?
Ms. Perlmutter. With that rule, in particular, we had
originally asked for comment because we had heard different
views expressed on the issue of the late fees. Once we looked
at the comments and evaluated the statute further, we concluded
that the statute was clear on the late fee issue and that it
would be up to the CRB to determine any different treatment of
late fees.
Mr. Fitzgerald. Very good.
Music fans, generally, want to listen to their favorite
artists and songs regardless of which record label or artist it
belongs to, or which PRO has particular rights to any given
song. We had a wonderful field hearing in Nashville in which
that point was brought up many times by many different
individuals.
They made the point; an individual song may have multiple
writers who belong to different PROs. So, radio stations and
music service providers generally need to license from all
PROs, rather than just one or two. This arrangement, as you
know, leads to a consistent upward trajectory of royalty rates
because, if a PRO, like a GMR, it gets a high rate through
contracting with in-demand songwriters, this rate will serve as
a benchmark for negotiations with other services for a new,
higher rate which can be used by the next PRO to negotiate a
new rate, and so on.
Can you talk about what your office is doing to keep the
music affordable to customers?
Ms. Perlmutter. The Copyright Office is not involved in
rate-setting, either through the PROs or in terms of the work
of the CRB in setting rates for music. So, we have a defined
scope of authority in this area, and there are certain things
we are regulating, but the actual amounts of the license fees
we do not handle.
Mr. Fitzgerald. So, would it be accurate to say, then, no
matter what happens with those rates, that you would not
intervene, and you certainly would not get involved in that
discussion?
Ms. Perlmutter. We would not intervene, although, to the
extent that Congress is looking at any legislative changes, we
would be happy to offer any kind of technical advice and
assistance.
Mr. Fitzgerald. OK. All 10 publishers represented by voting
members of the MLC's Board are also represented on the Board of
the National Music Publishers' Association. Are you at all
concerned that this overlap raises questions about the MLC's
independence and the ability to function as a neutral
administrator?
Ms. Perlmutter. We're aware that some have raised those
concerns, and we do think it would be valuable to address those
concerns and avoid any perceptions of a lack of balance. So, we
would certainly be happy to discuss further any potential
solutions. Any change in the board composition, however, would
require statutory change.
Mr. Fitzgerald. Just one final question. The other thing we
heard on the field hearing was that there were many artists
that the MLC was having a difficult time even locating. Have
you been involved in that process at all, or do you oversee
that in any way?
Ms. Perlmutter. Yes, and we did do a study and issue a
report about two years ago now with recommendations for the MLC
and how to improve matching, to be able to pay out to the
appropriate copyright owners.
Mr. Fitzgerald. Do you think that has improved or are you
tracking that at all?
Ms. Perlmutter. I think it's improving. In terms of the MLC
paying out under the new, blanket statutory license, I
understand they've reported about a 90 percent match rate at
this point. Then, in terms of matching the historical unmatched
royalties, we believe that's improving.
We recommended that the MLC not pay out for at least five
years, so that there was more time than the statute, the
statutory minimum to make, to do as much matching as possible
and identify as many copyright owners as possible. It is our
understanding that at this point in time the MLC is not
planning to pay out the money anytime in the near future.
Mr. Fitzgerald. Very good.
I yield back.
Mr. Issa. I thank the gentleman.
We now recognize the Ranking Member of the Subcommittee,
Mr. Johnson of Georgia.
Mr. Johnson of Georgia. Thank you, Mr. Chair.
Register Perlmutter, in its guidance, the Copyright Office
argued that both the Constitution and the Copyright Act define
an author as ``a human entity,'' excluding nonhuman creations,
and noted that this stance is reinforced by various court
rulings, and concluded that those using AI in their creations
can claim copyright protection only for their human
contributions.
How did the Copyright Office make that determination, and
how will that bright line help artists and creators survive in
fields that will increasingly be filled with AI-generated
works?
Ms. Perlmutter. Well, I would divide my answer into two
parts. So, the first part is the issue of the protectability of
the output, the protectability of individual works that are
generated using AI. To the extent that humans use AI as a tool,
that should not affect the protectability of the work. The
human author would still be the creator and the work would be
protected by copyright. Using AI as a tool is not something
entirely new. It has been used, in particular, in the music
field, for example, for decades in various ways.
The issue that we address because it was squarely put
before us, in particular, in the Thaler case, was a situation
where the claim was that a work was generated solely by a
computer and in that situation, looking at the word author and
both the Constitution and the statute, as you have noted, Mr.
Johnson, we believe that this term implies human authorship and
not that of a machine. That position was borne out by a number
of court cases over decades, if not centuries. So, we were
pleased to see the District Court and the District of Columbia,
as I mentioned, agree with us on that. What we are trying to do
is to guide people as to how they can register works that
include generative AI outputs. In that context, we said they
should just disclaim the portion of the work that was generated
by AI and we will issue a registration for the work as a whole
where there is some human authorship.
The second part of the question I think really relates to
what effect it will have on the ability of human creators to
make a living when their works are input into the computer for
purpose of machine learning and then the output might compete
in the marketplace with their work. That is an issue that we
are going to be studying that we have asked for comments on our
Notice of Inquiry. We agree, it is a critically important issue
for the future. We need to make sure that we do not inhibit the
development of very exciting new technology, while at the same
time ensuring that human creativity continues to thrive.
Mr. Johnson of Georgia. Thank you. Many AI models are
trained on copyrighted data, but not all the training data is
licensed. Why is it important that copyrighted information used
to train AI be licensed?
Ms. Perlmutter. One of the issues that is presented both
before us and that we have asked questions about in the Notice
of Inquiry and also in some of the court cases that are pending
is the extent to which any use of copyrighted works in training
the AI may qualify as fair use. To the extent it qualifies as
fair use, it would not need to be licensed. To the extent there
are certain uses that do not qualify as fair use, there would
be a need to obtain licenses. Then we have asked the further
question of how would that be done as a practical matter? Can
it be handled through direct licensing? Should it be handled
through collective licensing? Or should Congress consider some
sort of new compulsory license system?
So, a lot of open questions still and we will be exploring
all those as we review the responses to our Notice of Inquiry.
Mr. Johnson of Georgia. Would licensing be sufficient?
Ms. Perlmutter. It would certainly deal with the need to
have authorization if a use falls within the rights under the
Copyright Act and doesn't qualify as fair use. So, it would
deal with the authorization element. Presumably, it will also
deal with the compensation element and depending on the terms
of license, at the license, it could also require credit being
given to the creators.
Mr. Johnson of Georgia. Thank you. My time is about to
expire, so I will yield it back.
Mr. Issa. I thank you for the six-seconds. We now go to the
gentleman from Virginia, Mr. Cline.
Mr. Cline. Thank you, Mr. Chair. Register Perlmutter, thank
you for being here. You do a great job, and your 444 employees
are well served.
I want to go back to the subject of a hearing we had back
in July with the MCA and Section 1201 and right to repair. The
DMCA was introduced back in 1998 to address the evolving
relationship between copyright, the internet, and merging
technologies. Section 1201, as you know, prohibits the
circumvention of technological protection measures designed to
safeguard copyrighted works. Your office oversees a triennial
rulemaking process pursuant to the DMCA to grant specific
exemptions to 1201, balancing copyright protection with
technological advancement. Some have called for reform of this
process, particularly those who rely on exemptions to enable
third-party repair of devices with TPMs.
With respect to those issues, do you believe that Section
1201 should be modified statutorily, or do you think the
Copyright Office will make an authority sufficient to address
the needs of copyright holders and consenters?
Ms. Perlmutter. My answer would really be both. We are able
through rulemaking to address the right to repair issue and we
have done it in I think in the last three successive
rulemakings. On the other hand, that is not the most efficient
way to do it. It requires a lot of work from stakeholders and
from the Copyright Office every three years. For that reason,
in our report on Section 1201 a few years ago, we did recommend
a statutory permanent exemptions for the right to repair in
appropriate circumstances and we still believe that would be
advisable.
Mr. Cline. Can you talk about what are the most important
reasons that led to your office's decision to expand those
exemptions in that space?
Ms. Perlmutter. We did hear that there were many situations
where because copyrighted software is incorporated into devices
of various kinds, and because it is protected by encryption or
other technological measures, that the MCA's prohibition on
circumventing technological measures made it impossible for
people to get access for purposes of repair. So, for that
reason, in particular contexts which were medical devices,
consumer devices, and vehicles, we did put in place, we
recommended to the Librarian of Congress, and she put in place,
I should say, exemptions allowing circumvention in those
situations to make repairs.
We continue to believe that is a valuable tool for people
to have and therefore we would support addressing it in a
permanent way rather than through a triennial rulemaking.
Mr. Cline. So, in your view, this expansion could lead to a
proliferation of third-party repair shops and emergence of a
more competitive and robust repair industry?
Ms. Perlmutter. It may, yes.
Mr. Cline. Now, that you have recently initiated your ninth
triennial rulemaking process, we know that petitions for
renewing existing exemptions and new exemptions were due August
25th, can you offer any insight into new exemptions that are
being requested and how they compared with requests over the
past several rulemakings?
Ms. Perlmutter. We have had a few new requests, not a
deluge of them. We never know until we get them. We will be
issuing a Notice of Proposed Rulemaking in October which will
set out our recommendations on all the requests, but I don't
think anything will be a tremendous surprise. There are a
number of requests to renew existing exemptions or expand them
and just a few new ones, not a lot.
Mr. Cline. Talking about TPMs, rights holders argue that
aside from protecting IP TPMs, ensure device safety, security,
and reliability, what has your office found in terms of whether
the increased granting of exemptions weaken IP rights for
owners and present security issues for consumers?
Ms. Perlmutter. We have not been made aware of any serious
problems arising from any of the exemptions that the Librarian
has granted. So far, what we have recommended, and the
Librarian has adopted, have been very carefully delineated
exceptions that are drafted in such a way that we believe
avoids negative impacts.
Mr. Cline. Thank you. I yield back.
Mr. Issa. I thank the gentleman. We now go to the Ranking
Member of the Full Committee, Mr. Nadler, for five minutes.
Mr. Nadler. Thank you, Mr. Chair, and thank you for
appearing before the Subcommittee. Unlike most other regulatory
agencies, the Copyright Office has taken concrete steps toward
studying and guiding the adoption of artificial intelligence
with its guidance and Notice of Inquiry. How do you see the
Copyright Office's role in AI regulation?
Ms. Perlmutter. We have two separate roles. One is our role
as the administrator of the Copyright Act and registering
applications or registration, accepting applications for
registration. There, we are having to make day-to-day decisions
looking at specific examples of works created using AI
technology and the extent to which there is a human
contribution that can be protected by copyright. So, in a way,
we are a bit of a natural laboratory for looking at what the
dividing line is between human and computer creation.
The second role that we play is as an advisor to this body
and also to the courts and to other Executive Branch agencies
on copyright law and policy. In our Notice of Inquiry, we are
looking at all these issues. We are looking at whether we need
to refine in any way the guidance that we issued in March on
registration and we are looking at what the policy implications
are and whether we would recommend any changes to legislation
or regulation.
Mr. Nadler. Well, you may have answered part of my next
question. How should intellectual property be considered when
Congress or the Executive Branch as a whole decides to regulate
AI?
Ms. Perlmutter. I am sorry. Can you repeat the question?
Mr. Nadler. How should intellectual property be considered
when Congress or the Executive Branch as a whole decides to
regulate AI?
Ms. Perlmutter. I think as a result of the input that we
will get in response to the Notice of Inquiry we will have a
better sense of the extent to which any changes to the
copyright law are necessary. I know that this overlaps a bit
with some of the questions more broadly affecting society from
the development, in particular, generative AI dealing with
security issues, and dealing with privacy issues. So, I think
it is important to keep an eye on potential overlaps and
potential relationships. At the moment, it appears that the
copyright issues could be dealt with separately, that we don't
yet see any necessary overlap between other initiatives
relating to AI, but that may change as we continue to examine
and follow what is happening in the space.
Mr. Nadler. Thank you. What are the potential pitfalls of
AI adoption in artistic fields like music, visual arts, and
writing?
Ms. Perlmutter. There have been great concerns expressed by
creators and performers about the impact of a proliferation of
AI-generated content on their ability to make a living. There
are still a lot of unknowns about what the economic effect will
be, and we are looking at those issues as well as the legal
issues.
In terms of the protectability of AI-generated content,
there are a lot of questions about the extent to which it might
replace human-generated content, what consumers are going to be
interested in seeing, hearing, and listening to. So, all these
issues are very much up in the air. We are looking to find out
through the Notice of Inquiry all the concerns that people
have, and we will take those into account in writing a report
once we have had a chance to review the input.
Mr. Nadler. Thank you. A professional portrait artist takes
an average of 3-6 months to create a work of art. As they
become more popular, their work will often change hands,
becoming more valuable with each sale. If the artist will never
see a painting of these later sales, does the Copyright Office
continue to support the passage of Resale Royalty Scheme for
visual artists?
Ms. Perlmutter. Yes, we do, for exactly the reason that you
mentioned. I think the copyright system for a long time has not
served visual artists who create unique works of art as well as
it has served other types of creators and having the ability to
obtain a share in the money that is made when an original work
is resold is one way to allow an artist to continue to benefit
from an increased demand.
Mr. Nadler. Thank you. My last question is under the
current statutory framework, radio broadcasters do not need to
compensate artists when their songs are played on AM/FM radio.
What would be the benefits of legislation requiring royalty
payments to the broadcast of artists' songs and what resources
would the Copyright Office need to distribute those royalties?
Ms. Perlmutter. This is an area where the Copyright Office
for many years has recommended legislation. The United States
is one of very, very few countries in the world that do not
provide a full public performance right for sound recordings
covering over-the-air broadcasts and we believe it is past time
to do so. If we do so, that will mean more money coming into
the United States from other countries that provide that right
to performers, but only on a reciprocal basis, so they will not
pay performers in the United States because the U.S. doesn't
have an equivalent rate.
Mr. Nadler. Thank you for supporting Chair Issa and my
legislation and with that, I yield back.
Mr. Issa. Thank you. We now go to the gentleman from Texas,
Mr. Moran, for five minutes.
Mr. Moran. Thank you, Mr. Chair. Director Perlmutter, thank
you so much for your time today. I wanted to start out by
talking a little bit about what kind of cooperation we have had
across the globe with dealing with copyrightability of AI-
generated works. Can you tell me what level of cooperation we
are currently engaging with other countries across the world
with respect to this issue?
Ms. Perlmutter. Yes, that is a very important question.
Thank you. Of course, we can't only look at the issues within
U.S. borders because the technology involved in AI can be
international in scale. The databases can include copyrighted
works from around the world. The technology is developed by
different groups around the world, so we need to think about
the international implications.
We are in touch with our counterparts in other countries. I
have been to numerous meetings just over the last nine months
with copyright policymakers in a plethora of different
countries, so we are trying to make sure that we talk to each
other, that we understand what is happening and why.
We are also participating in conversations through the
World Intellectual Property Organization with other countries
as well. The goal really is to say having a certain level of
consistency and how we treat this issue and how we address the
issue is going to be desirable. We don't need to have identical
laws, but we should have consistent laws.
Mr. Moran. I like the way you put that because we certainly
want to contain our sovereignty here and make our own
determinations, but consistency certainly provides some
benefits to our human artists.
Who is leading on this issue as you have been talking
across the globe? I met with some folks from the E.U.
Parliament last week. They seemed to be doing a lot in this
space to determine what kind of framework should exist in the
copyright and AI space. Who would you say is somebody we need
to look at for leadership in this world that we need to
cooperate with?
Ms. Perlmutter. It is interesting because the Copyright
Office here has become the leader in looking at the
copyrightability issue because of our system of examining works
submitted for registration. So, a lot of countries are looking
to us on that particular issue.
On other issues, yes, the E.U. has moved forward with The
AI Act, which I believe is expected to go into force later this
year, and that deals with transparency issues and the question
of providing information about what copyrighted works have been
input into the machine. Then, we have seen a lot of activity
also in other countries, such as, Japan, Korea, and Singapore,
for example.
Mr. Moran. Let me switch gears and talk about the Copyright
Claims Board for a moment. At the one-year mark of its
establishment, there were over 500 cases filed with the
Copyright Claims Board. Does the Copyright Office have enough
resources to ensure the smooth operation of that board? Could
you talk about that?
Ms. Perlmutter. Yes. Thank you for that question. So far,
the answer is yes, but we need to see how this evolves in the
coming months and years. So far, the cases are coming in at a
pretty steady rate and our fears in the beginning that we might
be inundated have not yet happened, so there is plenty of work.
At the moment, we have a team that is large enough to handle
it. We will certainly come back and report if that turns out
not to be the case.
Mr. Moran. Speaking of reports, how are you keeping track
of the statistics and the outcomes of the CCB cases? Talk to me
about that.
Ms. Perlmutter. A lot of data and information is publicly
available on our website about the cases that have been filed.
We are also developing internal charts and graphs to try to get
a sense of what is happening and in what sectors. We will
continue to make information available as is feasible.
Mr. Moran. All right, and then finally, we have got about a
minute, I want to go back to this NOI that you were talking
about earlier. In August of this year, as you noted, ``the
Copyright Office initiated a Notice of Inquiry to better
understand the use of copyrighted works in training AI
models.'' As of about September, mid-September, I think you
guys have had 15,000 responses. I know we are not yet to the
deadline of responses. Can you highlight some of the key
feedback you have already received including before the NOI?
Ms. Perlmutter. I think what tends to happen is that we
will get a rush of early responses as soon as we publish our
notice and that is what happened. That is why the numbers seem
high. Then we will tend to get many of the much more detailed
responses late, like probably right near the deadline. So, at
present, we are still going through the responses we have
received so far. I think they have tended to be mostly
reactions to the fact that we are doing the study and people
talking about how important this is and how major the impact
could be on their lives and careers.
Mr. Moran. OK, thank you, Director. I yield back.
Mr. Issa. I think the gentleman. We now go to the
gentlelady from North Carolina, Ms. Ross.
Ms. Ross. Thank you, Mr. Chair. Thank you for joining us
today, Register Perlmutter.
My home district in North Carolina is home to creators in a
variety of creative fields from musicians to visual artists to
filmmakers to writers. As a matter of fact, we are having a
bluegrass festival, international bluegrass festival this
weekend which I am sorry I will not be there for. We all
benefit from the work of these artists who make our lives
fuller and richer with their creativity, their skill, their
dedication, and their craft. They would not be able to produce
the work they do without strong copyright protections that
allow them to make a living as creators. So, I am grateful for
this opportunity to hear from you about initiatives that the
Copyright Office is taking to ensure that creators can continue
to profit from their work.
In June 2022, the Copyright Office report titled,
``Copyright Protections for Press Publishers,'' in that report,
the Office addressed the fact that there is still no practical
method available from the Copyright Office for copyright owners
of dynamic and voluminous content, such as a news website or a
mobile app to register their content. The advent of AI which
enables new types of infringement puts publishers in an even
more precarious position, as we know the news industry is also
in a precarious position. In the report, the office stated the
office takes these concerns seriously and is considering how to
best address them as part of its ongoing modernization
initiative.
How has the office taken steps to address these concerns.
Ms. Perlmutter. We do take the need for this very seriously
and in fact, we are working on a Notice of Proposed Rulemaking
at this moment on a group option for registering news websites,
so that should appear soon. We believe we can do it within our
existing technological capabilities and not wait until ECS is
fully developed.
Ms. Ross. Thank you. I am the lead democratic sponsor,
along with the chair of a bill called the Pro Codes Act which
would protect the copyrights of standards and codes
incorporated into law while requiring a free version of each
code to be published online. At our July markup of the Pro
Codes Act, questions were raised about whether granting
copyrights in industry standards such as the National
Electrical Code was somehow improper or violated the government
edicts doctrine. Given that the Copyright Office has issued
thousands of copyright registrations for industry standards
over the course of decades, it appears that the Copyright
Office does consider such standards to be copyrightable.
Can you further explain why allowing the private sector
authors of standards to obtain and protect copyrights, even
where those standards are later incorporated by reference into
law does not violate the government edicts doctrine?
Ms. Perlmutter. Our view is that the privately authored
sets of standards are protected by copyright, as you note. This
is an important incentive for the work that is required to
develop these standards which can require a lot of thought and
expertise, but that at the same time, the public should have
access to them when they are incorporated into law because the
public does have a right of access to the law.
The courts have generally dealt with this issue so far by
saying that while the standards themselves may be protected by
copyright, the use of them generally falls under fair use if it
is for purposes of using and applying and understanding the
law. So, at present, we think the courts are handling this in
an appropriate way.
Ms. Ross. To compensate copyright holders for the use of
their work to train AI models, some have suggested creating a
licensing-type system for these works, similar to what the
Music Modernization Act streamlined for music. From what you
have seen of the Music Modernization Act effects, what do you
think about these proposals?
Ms. Perlmutter. First, I would say I think the Music
Modernization Act is working very effectively. The issue of
whether there should be a similar license for training AI is a
complicated one. It would involve all types of works, not just
musical works, and be far reaching in its scope. That is why we
are asking questions about it in our Notice of Inquiry. I think
there are a lot of practical issues involved that need to be
explored including about how the license fees would be set, how
it would be distributed, and if it is, how it can be made
feasible given the volume of works that would be involved.
Ms. Ross. Thank you, Mr. Chair. I yield back.
Mr. Issa. Thank you. Thank you very much for your line of
questioning.
Now, we will go to the gentlelady from Florida, Ms. Lee.
Ms. Lee. Good morning, and thank you so much for being
here. I would like to start by returning to the discussion of
the Copyright Claims Board and specifically, I think it is very
interesting that you made great use of that, and it seems to be
off to a productive start. I am interested though, there is a
set of claims that were rejected because they were
noncompliant, or they didn't meet the technical standards for
going through that resolution process.
Could you share with us what your office is doing to
provide resources or guidance for those who are trying to
utilize the Copyright Claims Board process?
Ms. Perlmutter. Yes, thank you for that question. I am very
proud of what we are doing to help people use the process. We
have a website with a wealth of information. We have a handbook
to tell people what they need to do to use the system. We have
people who answer the phone and answer questions if anyone
calls. We are also reaching out to get law school clinics and
volunteer lawyers for the arts signed up to help people
navigate the system even though they do not need to be
represented by an attorney. We are doing a lot of outreach, a
lot of public speaking about it.
The system is set up so that there are a number of
safeguards against inappropriate claims being brought so that
was very carefully thought through by Congress in enacting the
CASE Act, of course, and there are several opportunities along
the way if a claim is noncompliant, the Board attorneys will
tell the claimant and the claimant has two opportunities to
revise the claim to make it compliant. I will say some of the
lack of compliance we see is people sometimes bringing claims
that, for example, don't involve copyright law, maybe a patent
claim, or a claim against a foreign respondent which is not
permitted under the statute.
Ms. Lee. Based on what you have seen so far with the
inception of the program and its progress to date, what do you
anticipate being the future of the program? Do you anticipate
it is going to be broadly utilized? Any challenges or things
that you need Congress to do to help you succeed?
Ms. Perlmutter. It seems to be very successful so far. The
public reaction, the public reviews of what is going on have
been positive. The system is working. We have had a number of
final decisions, some based on settlements that have been
entered into the record and we have found that while some
respondents opt out, many respondents also are happy to proceed
in the CCB and prefer the certainty of knowing that their
potential damages are limited. So, we think it is working well
and the upward trend has been quite steady, so we think that it
will continue to go up, but presumably unless something really
unusual happens, the trajectory won't change. We may need
further resources, especially as we seek to further develop the
eCCB, which is the electronic case management system. We will
keep the Subcommittee apprised.
Ms. Lee. Speaking of your electronic case management
system, one of the things you touched on in your testimony was
your efforts to modernize and replace legacy IT systems. Would
you share with us what you are working on there and your
strategic plan to ensure that you are making the right
investments that are going to stand the test of time?
Ms. Perlmutter. It has been a huge priority for the
Copyright Office. IT modernization, we are now in the fifth
year of our originally planned modernization effort and now we
are moving toward continuous development. The main highlights,
as I described in my testimony, already we have moved from a
very archaic paper-based recordation system in the last two
years to one that is online. That has been huge. We are now
focusing on registration. We expect in the next year to begin
user testing for various components of the registration system
including the handling of electronic deposits.
We are also going to start recording Notices of Termination
online which we have not yet been able to do and that will
start in the next year. We are making more and more historical
records available online, so people no longer need to come to
our offices in Washington, DC, to do research. We are
experimenting with AI tools ourselves to extract metadata from
our records. So, we have a lot planned.
Ms. Lee. Thank you. I yield the balance of my time to the
Chair.
Mr. Issa. Thank you. That metadata brings up just one quick
question I will inject. You mentioned the disclaiming of AI
produced. Can you envision that along with disclaiming that
they include the metadata that shows where they--what they
ingested, how it turned into what they produced that they are
disclaiming? Are you considering that within your IT
modernization?
Ms. Perlmutter. That is a very interesting idea. We have
not yet considered it. We generally have been trying, however,
to make the application for registration as easy and simple as
possible and have heard some concern that we might be asking
for too many details and people would prefer that we make it
still easy to fill out an application without having to provide
a lot of information. We have to keep that in mind as we look
at that.
Mr. Issa. Thank you. To be continued. With that, I go to my
colleague and classmate from many years ago in Congress, Mr.
Schiff.
Mr. Schiff. Thank you, Mr. Chair. Thank you, Director, for
being here. Earlier this year, it was reported that TikTok was
limiting a number of songs available within its app for some
users in Australia in an attempt to test the importance of
music to the app's users. As a result, the number of people
using TikTok in Australia declined for three consecutive weeks
after the rollout of the test, according to numbers from the
data research firm data.ai.
The test revealed what many creatives already know to be
true. Much of TikTok's success can be attributed to copyright
works by musical artists, many of whom I represent in my Los
Angeles District, and who are often under compensated for the
use of their work on TikTok's platform.
Director, in light of what we have seen in Australia and
what is suggests, what thoughts do you and your office have
about how the United States can ensure that creators and rights
holders are protected when their work is used on digital
platforms?
Ms. Perlmutter. Well, this is obviously a top priority for
all of us in the copyright field. We have done a number of
things over the years to contribute to that effort.
We, of course, don't have enforcement capabilities in the
Copyright Office, but we do provide input from a policy
perspective on what the law should say. That has included over
the last few years a report on Section 512 of the DMCA and some
improvements that could be made in that section, as well as
looking at how copyright owners are using technological
measures to identify and protect their works. We continue to be
interested in that and to look at ways that we can convene
interested parties to continue to discuss improvements.
We also continue to follow and review what other countries
are doing in this respect and what tools have been adopted
elsewhere that could be useful to American right holders.
Finally, we work closely with the Executive Branch on
initiatives that have to do with international enforcement,
including the U.S. Trade Representatives Special 301 report and
Notorious Markets report.
Mr. Schiff. In terms of TikTok, do you think the remedy is
primarily an enforcement one, or are there legislative changes
you have recommended as a policy matter?
Ms. Perlmutter. We have not, to date, looked at potential
legislative changes. I have not been made aware of any
proposals from stakeholders but would be interested in hearing
more about it.
Mr. Schiff. Let me turn to something else. Earlier this
year the Supreme Court decided the Warhol v. Goldsmith case
dealing with fair use. How do you think that decision impacted
or clarified the way courts are supposed to apply the fair use
analysis? How do you expect this ruling to impact any future
rulemaking from the Copyright Office regarding the use of
copyrighted material to train AI models?
Ms. Perlmutter. We agree with the Supreme Court decision,
which adopted a lot of the analysis of the U.S. Government. The
government participated in that case as an amicus.
It may be that the case will have an influence on the fair
use analysis of the ingestion of copyrighted works for purposes
of machine learning. It requires, well, it affirms the need to
look at the markets for both the original work and the work
that is based on the original work, the work that is the
subject of the infringement claim, and to look at the extent to
which they are competitive, that they share the same market,
and the extent to which the defendant's use is commercial in
nature.
So, how that will apply to the analysis of fair use in the
ingestion of copyrighted content for AI training is still up in
the air. It will certainly affect it. We are watching the court
cases to see how the courts react. We will, we have asked
questions about this in our Notice of Inquiry. We will be
analyzing it once we receive the responses as well. It does
seem as if it will have an impact.
Mr. Schiff. Well, certainly by those measures, whether it
is in competition, the marketplace that test I think would be
met, and that it is economic in nature is also plainly the
case. Well, thank you, Director. I appreciate your work.
With that, I will yield back, Mr. Chair.
Mr. Issa. Could I ask you yield to me?
Mr. Schiff. Of course.
Mr. Issa. I just wanted to followup on one thing. You said
fair use and then said use in commerce. Would you like to
expand that linkage that fair use for not commerce is
dramatically different than fair use when it is, in fact,
turned into revenue?
Ms. Perlmutter. Well, fair use is very context specific and
requires weighing a lot of factors, as I am sure you're aware,
Mr. Chair. One of the factors are the nature of the use,
including the extent to which it is commercial or
noncommercial. So, the courts will look at that as part of the
total weighing, but it is not determinative. What the Supreme
Court did in the Warhol case was to elucidate a bit further
what the relevance was of the commerciality of the use.
Mr. Issa. Thank you. Thanks for expanding.
With that, we go to the gentleman from South Carolina, Mr.
Fry.
Mr. Fry. Thank you, Mr. Chair.
Thank you, Madam Register, for being here today. I really
appreciate the importance of this hearing and your testimony
today.
Judge Moran hit on this earlier. I want to explore this
topic a little bit further. In August of this year, the
Copyright Office initiated the NOI, the Notice of Inquiry, to
better understand the use of copyrighted works in training AI
models. Judge Moran asked, ``but I want to expand on that a
little bit.'' What is--and I understand the themes. I know that
the time is not yet finished in which people can comment. What
are some of the early themes that you are seeing or the
individual comments that come to your mind as some of the early
feedback that you are receiving on the NOI?
Ms. Perlmutter. I can't say that I have read all the
thousands of comments we have received so far. Lawyers in the
office are in the process of doing that. My understanding is
most of the early comments came from individual creators and
artists expressing concerns over what is happening.
Mr. Fry. Just about the use of AI within their field?
Ms. Perlmutter. Yes.
Mr. Fry. Are they making suggestions on how to fix that, or
they are just airing grievances?
Ms. Perlmutter. Let me get back to you with an answer to
that question. I will say it is interesting because creators
and artists are very much on both sides of the issue in various
ways, because I do want to make the point that many of them are
using AI as a tool in the creation process and want to be sure
they can continue to do that. At the same time, they are
concerned about what the impact will be if AI-generated content
is competing in the marketplace with their works.
Mr. Fry. Madam Register, do you have any idea on the timing
of when the Copyright Office might be able to make
recommendations to us or a report to us based on this feedback?
Ms. Perlmutter. I would like to say as soon as possible. I
know that timing is important here because technology is
evolving very rapidly and its impact is likely to start being
felt. We would like to be able to make recommendations in a
short timeframe.
At the same time, we are aware that we are likely to be
getting tens if not hundreds of thousands of comments. The
comment period will end at the end of November. We will need
some time to finish analyzing and absorbing and then writing
the report. Hopefully, it will certainly be in 2024. I
certainly hope it will be in the first half of 2024.
Mr. Fry. Thank you. Some of the most resource intensive
projects that the Copyright Office has on its agenda over the
next six months, I imagine that is probably one of them. Do you
have the resources to appropriately carry out those tasks,
taking into account stakeholder comments, where applicable, and
giving them due consideration?
Ms. Perlmutter. I really appreciate that question. I
believe we have the resources to do that at this point. It is
true that we find ourselves doing more and more work as the
copyright policy issues proliferate these days. At some point,
we may need further resources. At present, we are well equipped
to handle the work on our plate.
Mr. Fry. Thank you for that. Under the MMA, the Copyright
Office has an ongoing regulatory authority to promulgate
regulations, to implement the law, and ensure that proper
functioning of the Medical Licensing Collective in its duties,
including ensuring it acts as a neutral administrator serving
three primary sets of stakeholders, songwriters, publishers,
and digital service providers.
What steps can the Copyright Office take within its
existing authority to improve oversight and transparency of the
MLC, particularly, as redesignation approaches in 2025?
Ms. Perlmutter. We are taking a number of steps in that
respect. We hold regular meetings separately with each of the
interest groups, with the publishers, the songwriters, the
digital service providers, and the MLC itself. We have open
rulemakings. We continue to have an open door for people to
raise issues with us. We will, as you mentioned, commence a
public notice asking for input with respect to redesignation
starting early next year. We do think it is very important to
make sure that all sides are heard and feel that they are able
to participate in the process.
Mr. Fry. What are some of the issues that are brought up in
this, I guess, this sphere?
Ms. Perlmutter. Well, the issues are constantly ongoing.
Som there is, I think in general the process is working quite
well. We do hear questions about perceptions that the board is
not balanced because there are more members from the publishing
community than the songwriter community. That is something that
is set by statute. So, that would be an issue for Congress to
address.
Mr. Fry. OK. Final question. Given the increased usage of
name, image, and likeness in the outputs of generative AI, what
are your thoughts on how Congress can best understand and
legislate on this issue?
Ms. Perlmutter. This is also one of the issues that we are
asking for input on in our Notice of Inquiry. A few years ago,
the Copyright Office did a study on moral rights in the United
States and as part of that study suggested that Congress would
consider, might consider enacting a Federal right of publicity.
There are a number of State laws that could cover the imitation
of name, voice, and likeness that is currently taking place
using generative AI. It is a patchwork of different laws in
different States. So, the question is whether a Federal law
that might either preempt State laws or at least set a ceiling
or floor for what State laws can say would be desirable.
Mr. Fry. Thank you. I really appreciate that.
With that, Mr. Chair, I yield back.
Mr. Issa. Thank you.
With that, we go to the gentlelady from California, Ms.
Lofgren.
Mr. Lofgren. It is Lofgren.
Thank you very much. Thank you--
Mr. Issa. Sorry, Zoe.
Mr. Lofgren. Ms. Perlmutter. First, just a kudos for the
modernization efforts. I know we are not done yet, but we have
made great progress. I want to thank you and the Librarian as
well.
Just a note on music modernization, I thoroughly think that
was a triumph of collegiality and cooperation. There are a few
bumps in the road. Part of it may be our fault. In terms of the
songwriters, I am hoping that you can make an extra effort to
reach out to them because of the, what Congress did, because we
want to make sure everyone is heard. I know you believe in
that.
I want to talk about the right to repair. We haven't talked
that much about Section 1201 in terms of reform. I raise that
with some trepidation because people who have content worry,
and I understand that. If you have a software tool that
protects content, if it protects a movie, if it protects your
song, if it protects your book, if it protects your visual art,
you shouldn't be able to break that. So, I am not suggesting
that.
However, if you are using a software tool to protect
something that isn't content, which is what you have addressed,
that is a different situation. I am wondering, you can legally
fix your tractor now, but you can't get the software that
allows you to fix your tractor under 1201.
So, I am wondering if we could craft a narrow exception
that would not cause any concern among the content community
that is specifically, narrowly, and unequivocally only directed
toward the right to repair. Do you think that is possible to
do? If so, would you help us?
Ms. Perlmutter. Yes, I do think it is possible. In our
office's report a couple of years ago on Section 1201, we noted
that we had been recommending and the Librarian had been
issuing exemptions to permit repair in certain circumstances
and that it might make sense to have a permanent statutory
exemption, so that this didn't need to be revisited every three
years. We would still support that.
Mr. Lofgren. We do have the triannual process for 1201
exemptions. I want to thank you. When we did 1201, and I was
here in the Congress when we wrote it, we didn't intend to
allow cell phone companies to protect their monopolies. It was
about protecting content, not factors, cell phones, or other
equipment.
You have a lot of work to do. I appreciate that, that you
are not asking for more staff. If we were able to either
streamline the 1201 process when it relates to noncontent
issues or create a narrow exception that protects content,
wouldn't that help relieve the workload in your office?
Ms. Perlmutter. Yes, I think there are several areas where
1201 could be amended that would make the process easier. I
have to say I was around also when 1201 was written.
Mr. Lofgren. Yes.
Ms. Perlmutter. The idea of the amount of work it would
require terrified me at the time. We have learned how to handle
it. It is a lot of work.
Mr. Lofgren. In terms of the right to your person, you have
talked about that. It is really State law, publicity about
yourself. It is important that we take steps more urgently on
that given the artificial intelligence is already here. People
are going to need to take action to protect their appearance
and their personhood in AI.
We don't have a statute I think in mind yet. I would very
much welcome and ask for your assistance in crafting something
that would provide that protection, but more importantly, that
would think through how we might give people an easy avenue to
enforce that right, because AI is massive. It is here. It is
not going to be stopped. Yet, the potential for people to have
their personhood misused is real, already here. Having a right
and being able to enforce that right are two different things.
Do you have comments on that?
Ms. Perlmutter. We would be happy to assist in drafting.
Mr. Lofgren. I appreciate that very much. We will followup
with you.
Mr. Chair, I yield back and thank you.
Mr. Issa. I thank the gentlelady.
We now go to the other gentleperson from California, Mr.
Lieu.
Mr. Lieu. Thank you, Mr. Chair. I want to thank Chair Issa
and Ranking Member Johnson for holding this important hearing
on copyright issues and to you, Register Perlmutter, for being
here today and for your hard work and leadership.
I also want to align myself with the comments of
Congresswoman Zoe Lofgren in terms of the Music Modernization
Act, which is coming up on the five-year anniversary. I was
proud to have been a cosponsor of the Act. The MMA transformed
the way songwriters, music publishers, and tech platforms
operate under streaming models.
The creation of the Mechanical License Collective has
streamlined the administration, collection, and distribution of
mechanical royalties. To date, it has paid out more than 1.3
billion to songwriters and publishers. As Co-Chair of the
Songwriters Caucus, I want to thank you and your entire team
for continuing to work with the MLC in meeting its mandate to
serve songwriters and ensure that they are compensated.
So, I know you have answered a number of questions on fair
use. I just want to understand. You are currently investigating
the issue of whether it constitutes essentially fair use if a
large language AI model trains itself on copyrighted works. Is
that correct?
Ms. Perlmutter. Yes, it is part of the questions in our
Notice of Inquiry. Of course, it also will depend on the exact
circumstances. So, not every use in training will be the same
for fair use purposes.
Mr. Lieu. When do you expect that to be completed?
Ms. Perlmutter. Well, that will be part of our report or
reports coming out of the NOI, so as early as possible in 2024.
Mr. Lieu. OK. As part of the investigation, it is not only
whether a large language model, let's say it trains itself on
copyrighted Taylor Swift songs, but also you are looking at if
it outputs lyrics in the style of Taylor Swift songs. Are you
looking at whether that is fair use as well?
Ms. Perlmutter. Yes. Of course, it will also be, whatever
we say in our report will also be informed by any court
decisions that have come out by that point in time in the cases
that have been brought.
Mr. Lieu. You have also said that essentially perhaps it
should be essentially a Federal right for voice recordings
because right now audio is not copyrighted, correct?
Ms. Perlmutter. The style or sound of a voice is not
protected by copyright.
Mr. Lieu. Your office had issued a report basically
suggesting that there should be some sort of Federal name,
image, and likeness law essentially.
Ms. Perlmutter. We recommended that Congress consider that,
yes.
Mr. Lieu. OK. That report laid out essentially a proposal
that was around the same level as what California's protections
are. Is that correct?
Ms. Perlmutter. I am not sure if I would say it was the
same level. It contained some of, it talked about the elements
of that legislation in California. Of course, the report came
out before we had the generative AI that we have today. So, any
legislation would need to take into account what the current
capabilities and contexts are.
Mr. Lieu. All right. Thank you. In March of this year, your
office issued, basically announced that a work that involves
``sufficient human authorship'' can be copyrightable. So, let's
say someone uses AI to write a song of which over half the
lyrics are written by AI. Would that constitute sufficient
human authorship? How do you determine what that means?
Ms. Perlmutter. I wouldn't impose any strict percentage or
numerical limit. So, less than half could be sufficient human
authorship. It is really decided on a case-by-case basis. I
find it very useful to think of this by analogizing it to a
human and human situation. So, for example, if two people
collaborated in writing a song, did each of them contribute
enough authorship to make them co-author of the song? If the
answer is yes in that context, then it should be the same in
the AI context.
Mr. Lieu. Ultimately for this you are just going to have to
rely on trust, right, that someone actually discloses they use
AI to help them with whatever creative work they are trying to
copyright.
Ms. Perlmutter. Yes, but that is true of all applications
we receive. Applicants have to certify that what they are
saying is truthful. It is a government document. There are
penalties for making false statements. Of course, you could
risk losing your registration.
Mr. Lieu. Do you view AI software as different than other
kinds of software that creators have been using for decades to
make their creations better?
Ms. Perlmutter. Yes, that is a good question. I think what
we are seeing with the new generative AI in the last nine
months or so does seem to be different, not just in speed but
in kind from the technologies that were used before. I think in
the past AI was used more as a tool. Now, what we see is that
it is generating content that if it were generated by a human
being would be protected by copyright. So, I think we have
moved to a somewhat different world than we were in before.
Mr. Lieu. Thank you.
I yield back.
Mr. Issa. I thank the gentleman.
We now go to the gentlelady from Pennsylvania, Ms. Dean.
Ms. Dean. I thank you, Chair. I thank the Ranking Member
for hosting and holding this oversight hearing. Thank you,
Register Perlmutter, for your work and the work of your 444-
person team to carry out the mission of the U.S. Copyright
Office, which is promoting creativity and free expression by
administering the Nation's copyright laws and by providing
impartial expert advice. I thank you for that work.
As an author myself, I benefit from and importantly rely
upon the protections that copyright law provides. As we have
all been discussing here, with the rise of AI these protections
are more important than ever. So, I guess my theme for my five
minutes is twofold. We must protect the spark of ingenuity that
makes copyrighted work uniquely human, and so at this
precipice, what are the top things we do to protect human
content.
So, I was thinking about your office. Are you able to
identify AI-generated work? Are you concerned about applicants
hiding AI copyright, AI work within a copyright registration
request application? What are the steps that you go through? I
have to say nightmarishly I go back to the days when I was a
professor of writing. I would be reading papers and wondering
are those student's words or are they someone else's. So, what
is it that you have to do now in this world to identify AI
within an application?
Ms. Perlmutter. Our examiners look carefully at an
application. We have issued this guidance that tells people if
AI-generated content is included they need to disclose it. You
are right. We have to trust, and we do take as factual the
statements that are made in an application unless we have some
reason to think they are not accurate.
So, there have been situations where it has been obvious
from the face of an application that some of the content in the
work may have been generated by AI. So, our examiners will go
back and ask the applicant about it and engage in some
correspondence. What we have seen so far in some of the cases
we have had is that the result is that the applicant will then
end up disclaiming some portion of the work that is submitted.
So that has worked, for example, in books where the
illustration was generated by AI, but the text and the
selection and arrangement of the text with the images was done
by the human applicant.
So, there is no way to be 100 percent sure. We are not
unique in having that problem in today's world. We do our best
to see from the face of the application and the deposit whether
we think there is AI-generated content. Again, there are some
legal requirements for people to be honest in the content, in
the statements that they make to a government office.
Ms. Dean. Absolutely. I was thinking of Mr. Lieu's question
and sort of the reverse side of that. The office, your office
instructs applicants to disclose all AI-generated content that
is more than de minimis. What is de minimis?
Ms. Perlmutter. It is not a new concept. So, the same rules
apply, for example, if the work contains content created by a
third party. That also has to be disclaimed if it is more than
de minimis.
We had a Webinar, we held a Webinar this spring where we
walked people through the guidance and gave them a number of
examples and told people that essentially what we mean by de
minimis is, more than de minimis is if the material that is
generated by AI would be protected by copyright if it were
generated by a human being, then it should be disclaimed.
Ms. Dean. Interesting, yes.
Ms. Perlmutter. So, we are not talking about just some
small portion or something incidental, but something that,
again, if a third party had generated it, if a human being had
generated it, you would disclaim it as a separately
copyrightable work.
Ms. Dean. That is logical. That makes sense.
Finally, in the few seconds I have left, what has your
office thought about or learned as a result of the labor
disputes with the writers and the studios? Full disclosure, my
son was on strike. He is part of the Writers Guild. In terms
of, obviously one of the big issues was use of AI in written
materials and in writers rooms. How is the Copyright Office
connected there? What clear rules should we be developing?
Ms. Perlmutter. Well, it is a good question. Clearly the
strike, both strikes, SAG-AFTRA and the Writers Guild, showed
how important AI has become to authors and performers, because
it was a critical part as I understand it of the negotiations.
We are now looking at what came out of the settlement on AI
and analyzing it. Of course, we don't play a role because these
are private party negotiations. It is not really a question of
what the law says. It is a question of what they agree to. It
is helpful to see what was thought to be important and what was
thought to be reasonable as the private parties were
negotiating. So, we will be looking carefully at the result.
Ms. Dean. Again, I thank you for all the work that you and
your office does.
I yield back.
Mr. Issa. I thank the gentlelady.
We now go to the gentleman from California, Mr. Kiley.
Mr. Kiley. I am going to yield the balance of my time. I
did just want to say thank you for your testimony. I know that
your office has been very active in getting input on these very
novel and vexing issues related to AI. I have certainly heard
from a lot of creators who have concerns in this area. I think
it is really important that we be mindful of the ability of
creators to protect their rights and intellectual property
moving forward.
So, with that, I yield to the Chair.
Mr. Issa. I thank the gentleman.
I am going to try and use up a little bit of Mr. Kiley's
time to challenge you to expand on making the record clear as
much as we can.
So, let me go through some things that might happen to
somebody's writings. You tell me whether they are clearly
disclaimable, clearly ineligible, or clearly not a factor, de
minimis as was said earlier. Spell check.
Ms. Perlmutter. No need to disclaim. That is a tool that
has been commonly used for a long time.
Mr. Issa. Word substitution to create enhancements such as
sort of a thesaurus might do.
Ms. Perlmutter. I would say the same thing.
Mr. Issa. Phrase enhancement, how is it better said.
Ms. Perlmutter. I would also just say the use of a tool to
enhance something that is created by a human would not raise an
issue as to copyrightability.
Mr. Issa. Enhancement of desirability based on machine-
learned desirability and focus group analysis.
Ms. Perlmutter. I also would see that more as a tool.
Again, the machine is not generating something that would be
copyrightable if a human generated it.
Mr. Issa. Final question in this series. Using these types
of tools but a regenerative AI, if you ingested an entire book
and 40 percent of the words, phrases, and content were changed
while, in fact, the underlying intent of the author wasn't and
that, but enhanced the desirability and all those other
phrases, would that, in fact, be either disclaimable or
ineligible?
Ms. Perlmutter. So, the applicant wrote a book, and then 40
percent of the book was--
Mr. Issa. Ingested, has the book evaluated by AI for all
these and other ideas, and the book, although the same book
with the same happy ending or unhappy ending was modified with
40 percent of the words and/or phrases being changed. I am
using 40 as an arbitrary number.
Ms. Perlmutter. It is a tougher question. I think what we
are talking about here is you have got an original
copyrightable work by a human, which is the book, and then a
derivative work created by the machine, which probably would
not be copyrightable as a separate work.
Mr. Issa. If you publish only the one, which is--and I am
using the example for a reason. Ms. Dean said it very well. You
are a professor. You are evaluating somebody's work. In my day,
I had to put up with the nuns in my college objecting to my
handwriting and my spelling. That has now been collectively,
creatively enhanced in a way that I would probably do better in
at least theology. However, I also wrote a book. I probably
would have written a much better book if it was ingested by
regenerative AI today.
The question is if Ms. Dean's book and my book were thrown
in and hypothetically 40 percent of the words and phrases were,
according to the machine, enhanced, but, in fact, the book was
still a true copy of the author's work, what do you disclaim?
What is copyrightable? What is not? It is an arbitrary
question, but I thought for the record if you would give us
your best analysis.
Ms. Perlmutter. Yes, it is a tough question. I think one
thing that is important to understand is that what we require
in a disclaimer is just something that says the work
incorporates AI-generated content and an explanation of what
the human contribution was. So, it doesn't require really fine
tuning and specificity.
I think what your question really goes to is that these are
not easy decisions. They are not easy lines to draw. So, one of
the things when I talk about the Copyright Office being a
natural laboratory for this is we are looking at individual
applications and making case by case determinations, which is,
of course, something that copyright law always requires. How do
you draw the line between an idea and expression? How do you
determine whether someone has contributed enough to be a co-
author?
We are learning more and more each day as we look at more
and more of these applications and make decisions on them. Some
of those decisions are being challenged in the courts and will
be being reviewed.
Mr. Issa. I thank the gentlelady. I thank the gentleman for
yielding the time.
We now go to Mr. Ivey for five minutes.
Mr. Ivey. Thank you, Mr. Chair. I want to commend you for
having this hearing. I have only been on the Committee for
about nine months now. I think we have done, of all the hearing
we have done, I think we have done four that I would put in the
substantive category. Three of them have been yours. I greatly
appreciate that fact.
Thank you, ma'am, for coming today. I appreciate the great
work you are doing.
I think we are in the hair-splitting segment of the hearing
now. I am going to followup on some of the questions the Chair.
I am trying to figure this out. I am a lawyer by training, but
not productive law. I did litigation, unlike the copyright
world where you are helping actual creations being made and
protected.
One of the things I was wondering about was, and I think
you mentioned this earlier. So, you have got sort of a human
creator who generates let's say a song and uses AI to help with
the--there is a portion of the song that is in part generated
by the AI. That is AI that has been trained and ingested other
information or music from other sources. I think what you were
saying was that the creator, when the human creator comes to
seek a copyright, has to explain or disclaim that there was AI
generation as part of the final overall work. Is that right?
Ms. Perlmutter. Yes, they would not have to disclaim the
use of AI as the tool, like to provide a beat or that kind of
thing or to change a rhythm. They would if there was let's say
a melody created entirely by AI that they incorporated into
their larger musical work.
Mr. Ivey. OK. So, we will go forward with that. I have got
a son that creates beats. He would disagree with where you drew
that line right there, but we will set that to the side.
So, the melody line, so, and let's say then that, because
we have got people who approached me, for example, and they
were complaining about AI being trained using their creative
work product. So, in the song that you and I just made, and
this author says I had AI-trained assistance in generating this
song. How was, or has there been a determination that has been
made with respect to the people who helped create the content
that was used to train the AI that was then used in this song?
How are they, or compensated for their work, or is it
recognized even?
Ms. Perlmutter. Yes, well, that is a critical question
right now. I know--and by the way I am a reformed litigator
myself. I know there are a number of cases pending in the
courts asking exactly that question.
So, assuming that the technology requires the reproduction
of the works in the computer for purposes of the machine
learning, there would be a prima facie case of infringement.
Then the question would be whether it is fair use. Whether it
is fair use might depend both on how it is used in the training
and also what the output is going to be and the extent to which
it competes in the market with the original.
Mr. Ivey. To what extent is this all going to be reliant on
I think your term was honesty, I'll say honor code, but, the
individual who--and I will preface the question with this just
in my experience, to the extent compensation and the dollars
get bigger, the honesty and honor code gets sometimes
marginalized.
So, you mentioned some of this before, which is the
certification gets signed and the like. Are there stronger
enforcement and protections that could be out there for the
creators of the original content that was ingested by the AI
model?
Ms. Perlmutter. Yes, well, the honor issue and the accuracy
and statements made to the government agency has to do more
with the applications for registration.
In terms of litigation, what is interesting is I am seeing
that more and more some of the companies that are producing and
distributing AI technology and tools are beginning to say that
they will license the copyrighted content that they use. Some
companies that create or license themselves copyrighted content
are saying that they are now going to create AI using that
licensed content. So, we may be getting more and more people
who are interested in making sure that there are no legal
questions surrounding their use of copyrighted works to train
the AI.
Other than that, the way to keep people honest I suppose is
these lawsuits that are being brought, some of them as class
actions. We will see what the courts have to say about the
legality of proceeding without a license.
Mr. Ivey. Spoken like a true litigator, right.
Then the last question in relation to that, the CCB I guess
is sort of an alternative approach. So, instead of going to
court, they can choose this path, but it is voluntary. I did
have a question about copyright trolls. Also, on the point you
just made, to what extent could that be an avenue where these
sorts of disputes get hashed out in a way that--again, a lot of
the people who are musicians or writing songs are, oh, jeez, I
am over my time, I apologize, Mr. Chair, aren't wealthy people.
Mr. Issa. You are the second last. So, you are only holding
up my questions. You are good.
Mr. Ivey. Thank you, Mr. Chair. I appreciate it.
Their ability to hire litigators can be limited. Is the CCB
an option or an avenue that might be a viable approach for them
or not so much?
Ms. Perlmutter. The CCB would be a possible approach. The
limitation is that they can't issue injunctions. So, they can't
order anyone to stop infringing. There is a limit of $30,000 in
damages.
So, certainly, the other option and the one that has been
pursued so far has been bringing cases in Federal court where
presumably, especially if it is a class action, there are
lawyers who will get compensated at the end of the day if the
action is successful.
So, I have been wondering whether we will start to see many
cases involving generative AI in the CCB. So, far that has not
begun to happen.
Mr. Ivey. Thank you, ma'am. Keep up the good work.
Thank you, Mr. Chair, for your indulgence.
Mr. Issa. Thank you. Thank you for your question, because I
am going to followup directly on that.
The CCB, as you say, voluntary, $30,000 limit, creation of
Congress, still in the early stages. However, the gentleman
from Maryland mentioned the word copyright trolls. We are
seeing a growth of what many would call copyright trolls,
people who gather a few earlier words, and even when they are
dramatically different, they still make the claim. There is a
high cost in Federal court.
As you know, your experience in the past at the Patent and
Trademark Office we created PTAB, and it is an alternate. It
serves a similar purpose of adjudicating and is appealable to
the Federal Circuit. It also, though, has a significant role in
changing, if items end up in the Federal court, how the Federal
court looks at them, because they give the obvious deference to
the decision of PTAB.
Although it is not universally loved here on the dais, in
this case, if we were to take the model in its infancy of CCB
but empower it to be a right to go and, in fact, the decisions
of it be if not substantive enough to be automatically
considered adjudication, if we considered them to be admissible
and some deference by the Federal court, would that be in your
opinion a helpful use and expansion of the role of that?
The reason I ask it is Federal judges are perplexed trying
to understand if this piece of a lyric should or shouldn't.
They generally get one case in a career, where you have 500
cases in front of you today.
Ms. Perlmutter. It is an interesting question. I hadn't
thought about that before. I will say, again, it is early days.
We have only had one full written decision on the merits so
far. We do have a very good board with three officers who are
extremely experienced with litigation on both sides of the
copyright issues. I have a lot of confidence in the quality of
the decisions that are going to be issued.
So, yes, it could be that they could be useful as
precedent. They certainly would be available to the public to
look at. I think that is a very important aspect of the CCB.
Mr. Issa. One of the reasons that I ask that question is
that I am an old Oversight and Reform Chair. The reforms that
are probably needed going forward to the institution of
Congress for which you oversee it really do have to do with how
the Patent and Trademark Office has been able to set fees, fund
itself, and expand to serve the very people whose intellectual
property is being protected. You have not been. You are a much
smaller entity, in no small part because you depend on Congress
to fund you.
So, one of the reforms that I want to put out here today
during your visit is the idea that we have the ability on this
Committee to change the structure, to provide for fee setting,
to provide for a level of autonomy and expansion to meet the
demands of, in fact, those entities you serve. I want to make
sure that we make it clear that is something we would like to
have an ongoing dialog about.
I will stop you because I don't want to be too long in this
final one.
Earlier today we talked about what I call standard
essential copyright, in other words, copyrights that are put
into products or even into standards. Would it be helpful for
Congress to, in fact, as we pretty well define standard
essential patents as something which does not give up the right
of the patent holder but gives a requirement to make available,
even at a fee but still available? Should copyrights that are
embedded in products, whether it is inside the deep bowels of
an automobile or the toner cartridge of a photocopier, should
that, in fact, be something that Congress addresses and
defines?
Ms. Perlmutter. I think that is an area where attention
could be useful. It may be that in the process of putting a
permanent exemption into Section 1201, which as I have said we
support, we could look at how to define the scope of that
exception.
Mr. Issa. Thank you. Last, copyright is a little different
than patents in that just because something has been
copyrighted 100 years ago or 1,000 years ago it doesn't take a
lot to, in fact, create a new copyright. Would you agree?
Ms. Perlmutter. Yes.
Mr. Issa. However, your office and the offices around the
world today do not maintain a single database so that one can
at least reference the scope and the prior art, if you will.
Do you believe that in the 21st century we should establish
a global copyright database, meaning that all the entities like
you around the world should, in fact, have an interoperable
database so that when there is a question of whether something
is new and original, or simply lifted as the gentlelady from
Pennsylvania implied as a former professor, in fact, that there
be an ever-expanding database? For lack of that, is that level
of trust on the author perhaps a little greater than it should
be in the 21st century?
Ms. Perlmutter. There is a lot to unpack in that question.
So, I would say, first, more data being available to the
public, including internationally, is definitely a positive
thing. It is certainly something that people have long
advocated for in the music space, because it would be very
useful given the complexity of music licensing to have that
kind of complete database.
One question is the extent to which that should be a
function of government and the extent to which it should be a
privately run database, which might be a little bit more
flexible and easy to keep up to date.
One of the issues with having some kind of international
government-run database is that, of course, copyright
registration is not mandatory. Not all works are registered.
So, by definition, it won't be completely comprehensive.
Another problem is that many countries around the world don't
have a registration system. So, we probably have the most
extensive registration system that exists. So, there are a lot
of issues to be dealt with.
Then, of course, it is also the case that, unlike patent
law, with copyright law the fact that someone else came up with
something similar in the past doesn't mean that your use is
infringing. It has to actually be copied from the original. So,
the mere fact that something was in a database, and you did
something similar would not necessarily establish infringement.
So, in a sense, there is less of a need on the copyright side
than there is in industrial property.
Mr. Issa. In closing, in discussions with many of the, or
three of the regenerative AI producers, they all tell me that,
in fact, providing the metadata that goes with the portion of
the ingested material, which, if you will, the AI brain used,
is, in fact, relatively easy if written into the code.
Should, like Europe, as we address mandates, should, in
fact, we ensure that this information is written into the code
so that, in fact, for you and other users that metadata is
available, and by the way, yes, potentially for litigators?
Ms. Perlmutter. This is definitely one of the issues we
heard a lot about in our listening sessions earlier this year
and something that we are asking about in the NOI. I think
there is a strong argument that some level of transparency is
important for many reasons, as you say, including for the
ability of copyright owners to know that their works were
incorporated in some way in the machine learning.
Mr. Issa. Thank you. Last, would you agree to take
questions for the record if they come to you timely in the next
week or so?
Ms. Perlmutter. I would be happy to. Thank you.
Mr. Issa. Thank you.
With that, this concludes today's hearing. I want to thank
our witnesses for appearing before the Committee.
Without objection, all Members will have five legislative
days to submit additional written questions for the witness and
additional materials for the record.
Without objection, this hearing is adjourned.
[Whereupon, at 11:53 a.m., the Subcommittee was adjourned.]
All materials submitted for the record by Members of the
Subcommittee on Courts, Intellectual Property, and the Internet
can
be found at: https://docs.house.gov/Committee/Calendar/ByEvent
.aspx?EventID=116404.
[all]