Notice2024-22767
Rytr LLC; Analysis of Proposed Consent Order To Aid Public Comment
Primary source
Metadata and text below are from the Federal Register, a public-domain U.S. government work. Always verify the official published version before relying on it for any legal matter.
Published
October 3, 2024
Issuing agencies
Federal Trade Commission
Abstract
The consent agreement in this matter settles alleged violations of Federal law prohibiting unfair or deceptive acts or practices. The attached Analysis of Proposed Consent Order to Aid Public Comment describes both the allegations in the complaint and the terms of the consent order--embodied in the consent agreement--that would settle these allegations.
Full Text
<html>
<head>
<title>Federal Register, Volume 89 Issue 192 (Thursday, October 3, 2024)</title>
</head>
<body><pre>
[Federal Register Volume 89, Number 192 (Thursday, October 3, 2024)]
[Notices]
[Pages 80565-80572]
From the Federal Register Online via the Government Publishing Office [<a href="http://www.gpo.gov">www.gpo.gov</a>]
[FR Doc No: 2024-22767]
-----------------------------------------------------------------------
FEDERAL TRADE COMMISSION
[File No. 232 3052]
Rytr LLC; Analysis of Proposed Consent Order To Aid Public
Comment
AGENCY: Federal Trade Commission.
ACTION: Proposed consent agreement; request for comment.
-----------------------------------------------------------------------
SUMMARY: The consent agreement in this matter settles alleged
violations of Federal law prohibiting unfair or deceptive acts or
practices. The attached Analysis of Proposed Consent Order to Aid
Public Comment describes both the allegations in the complaint and the
terms of the consent order--embodied in the consent agreement--that
would settle these allegations.
DATES: Comments must be received on or before November 4, 2024.
ADDRESSES: Interested parties may file comments online or on paper by
following the instructions in the Request for Comment part of the
SUPPLEMENTARY INFORMATION section below. Please write ``Rytr LLC; File
No. 232 3052'' on your comment and file your comment online at <a href="https://www.regulations.gov">https://www.regulations.gov</a> by following the instructions on the web-based
form. If you prefer to file your comment on paper, please mail your
comment to the following address: Federal Trade Commission, Office of
the Secretary, 600 Pennsylvania Avenue NW, Mail Stop H-144 (Annex R),
Washington, DC 20580.
FOR FURTHER INFORMATION CONTACT: Division of Advertising Practices,
Bureau of Consumer Protection, Federal Trade Commission, 600
Pennsylvania Avenue NW, Washington, DC 20580.
SUPPLEMENTARY INFORMATION: Pursuant to section 6(f) of the Federal
Trade Commission Act, 15 U.S.C. 46(f), and FTC Rule Sec. 2.34, 16 CFR
2.34, notice is hereby given that the above-captioned consent agreement
containing a consent order to cease and desist, having been filed with
and accepted, subject to final approval, by the Commission, has been
placed on the public record for a period of 30 days. The following
Analysis to Aid Public Comment describes the terms of the consent
agreement and the allegations in the complaint. An electronic copy of
the full text of the consent agreement package can be obtained at
<a href="https://www.ftc.gov/news-events/commission-actions">https://www.ftc.gov/news-events/commission-actions</a>.
You can file a comment online or on paper. For the Commission to
consider your comment, we must receive it on or before November 4,
2024. Write ``Rytr LLC; File No. 232 3052'' on your comment. Your
comment--including your name and your State--will be placed on the
public record of this proceeding, including, to the extent practicable,
on the <a href="https://www.regulations.gov">https://www.regulations.gov</a> website.
Because of heightened security screening, postal mail addressed to
the Commission will be subject to delay. We strongly encourage you to
submit your comments online through the <a href="https://www.regulations.gov">https://www.regulations.gov</a>
website. If you prefer to file your comment on paper, write ``Rytr LLC;
File No. 232 3052'' on your comment and on the envelope, and mail your
comment to the following address: Federal Trade Commission, Office of
the Secretary, 600 Pennsylvania Avenue NW, Mail Stop H-144 (Annex R),
Washington, DC 20580.
Because your comment will be placed on the publicly accessible
website at <a href="https://www.regulations.gov">https://www.regulations.gov</a>, you are solely responsible for
making sure your comment does not include any sensitive or confidential
information. In particular, your comment should not include sensitive
personal information, such as your or anyone else's Social Security
number; date of birth; driver's license number or other State
identification number, or foreign country equivalent; passport number;
financial account number; or credit or debit card number. You are also
solely responsible for making sure your comment does not include
sensitive health information, such as medical records or other
individually identifiable health information. In addition, your comment
should not include any ``trade secret or any commercial or financial
information which . . . is privileged or confidential''--as provided by
section 6(f) of the FTC Act, 15 U.S.C. 46(f), and FTC Rule Sec.
4.10(a)(2), 16 CFR 4.10(a)(2)--including competitively sensitive
information such as costs, sales statistics, inventories, formulas,
patterns, devices, manufacturing processes, or customer names.
Comments containing material for which confidential treatment is
requested must be filed in paper form, must be clearly labeled
``Confidential,'' and must comply with FTC Rule Sec. 4.9(c). In
particular, the written request for confidential treatment that
[[Page 80566]]
accompanies the comment must include the factual and legal basis for
the request and must identify the specific portions of the comment to
be withheld from the public record. See FTC Rule Sec. 4.9(c). Your
comment will be kept confidential only if the General Counsel grants
your request in accordance with the law and the public interest. Once
your comment has been posted on the <a href="https://www.regulations.gov">https://www.regulations.gov</a>
website--as legally required by FTC Rule Sec. 4.9(b)--we cannot redact
or remove your comment from that website, unless you submit a
confidentiality request that meets the requirements for such treatment
under FTC Rule Sec. 4.9(c), and the General Counsel grants that
request.
Visit the FTC website at <a href="https://www.ftc.gov">https://www.ftc.gov</a> to read this document
and the news release describing the proposed settlement. The FTC Act
and other laws the Commission administers permit the collection of
public comments to consider and use in this proceeding, as appropriate.
The Commission will consider all timely and responsive public comments
it receives on or before November 4, 2024. For information on the
Commission's privacy policy, including routine uses permitted by the
Privacy Act, see <a href="https://www.ftc.gov/site-information/privacy-policy">https://www.ftc.gov/site-information/privacy-policy</a>.
Analysis of Proposed Consent Order To Aid Public Comment
The Federal Trade Commission (``Commission'') has accepted, subject
to final approval, an agreement containing a consent order from Rytr
LLC (``Rytr''). The proposed consent order (``proposed order'') has
been placed on the public record for 30 days for receipt of comments
from interested persons. Comments received during this period will
become part of the public record. After 30 days, the Commission will
again review the agreement and the comments received and will decide
whether it should withdraw from the agreement and take appropriate
action or make final the agreement's proposed order.
This matter involves Rytr's marketing and offering for sale an
artificial intelligence ``writing assistant'' service intended to
generate unlimited content for consumer and customer reviews and
testimonials. Subscribers to the Rytr service, which uses generative
artificial intelligence, can quickly and easily generate unlimited
written content for a variety of ``Use Cases.'' One of these Use Cases
was the ``Testimonial & Review'' Use Case, which Rytr began offering in
April 2021.
According to the Commission's proposed complaint, the review
generation service generated reviews that would almost certainly be
false for the users who copied them and published them online. In many
instances, the resulting reviews featured details that would deceive
potential consumers deciding to purchase the service or product
described. The proposed complaint asserts that at least some of Rytr's
subscribers utilized the service to produce hundreds and, in some
cases, thousands of reviews.
The proposed complaint alleges that Rytr provided the means and
instrumentalities to its users and subscribers to generate written
content for consumer reviews that was false and deceptive. The
complaint also alleges that Rytr engaged in an unfair business practice
by offering a service that was intended to quickly generate unlimited
content for consumer reviews and created false and deceptive written
content for consumer reviews.
The proposed order contains provisions designed to prevent Rytr
from engaging in these and similar acts and practices in the future.
Provision I bans Rytr from advertising, promoting, marketing, offering
for sale, or selling any service dedicated to or advertised, promoted,
or offered as generating consumer or customer reviews or testimonials.
Provisions II through VI of the proposed order contain reporting
and compliance provisions. Provision II mandates that Rytr acknowledge
receipt of the order, distribute the order to principals, officers, and
certain employees and agents, and obtain signed acknowledgments from
them. Provision III requires Rytr to submit compliance reports to the
Commission one year after the order's issuance and submit notifications
when certain events occur. Under Provision IV, Rytr must create certain
records for 20 years and retain them for five years. Provision V
provides for the FTC's continued compliance monitoring of Rytr's
activity during the order's effective dates. Finally, Provision VI
provides the effective dates of the order, including that, with
exceptions, the order will terminate in 20 years.
The purpose of this analysis is to facilitate public comment on the
proposed order. It is not intended to constitute an official
interpretation of the complaint or proposed order, or to modify in any
way the proposed order's terms.
By direction of the Commission, Commissioners Holyoak and
Ferguson dissenting.
Joel Christie,
Acting Secretary.
Dissenting Statement of Commissioner Melissa Holyoak Joined by
Commissioner Andrew N. Ferguson
As I have suggested recently in other contexts, the Commission
should steer clear of using settlements to advance claims or obtain
orders that a court is highly unlikely to credit or grant in
litigation.\1\ Outside that crucible, the Commission may more readily
advance questionable or misguided theories or cases.\2\ Nevertheless,
private parties track such settlements and, fearing future enforcement,
may alter how they act due to a complaint's statement of the alleged
facts, its articulation of the law, or how a settlement order
constrains a defendant's conduct.\3\ In all industries, but especially
evolving ones like artificial intelligence (AI), misguided enforcement
can harm consumers by stifling innovation and competition. I fear that
will happen after today's case, which is another effort by the Majority
to misapply the Commission's unfairness authority under section 5
beyond what the text authorizes. Relatedly, I believe the scope of
today's settlement is unwarranted based on the facts of this case. I
respectfully dissent.
---------------------------------------------------------------------------
\1\ See, e.g., Dissenting and Concurring Statement of
Commissioner Melissa Holyoak, In re Coulter Motor Company, LLC, FTC
No. 2223033, at 3 n.17 (Aug. 15, 2024) (``It is no coincidence that
the Commission has asserted its novel `unfair discrimination'
authority only outside the scrutiny of courts and in the context of
consent orders.''), <a href="https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-re-coulter-8-15-24.pdf">https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-re-coulter-8-15-24.pdf</a>; see also
Statement of Commissioner Melissa Holyoak, In re Asbury Automotive
Group--McDavid Group; Matter No. 2223135, at 1 n.1 (Aug. 16, 2024),
<a href="https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-re-asbury8-16-24.pdf">https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-re-asbury8-16-24.pdf</a>.
\2\ See Dissenting Statement of Commissioner William E. Kovacic,
In re Negotiated Data Solutions, LLC, File No. 051-0094, at 3 (Jan.
23, 2008) (``The prospect of a settlement can lead one to relax the
analytical standards that ordinarily would discipline the decision
to prosecute if the litigation of asserted claims was certain or
likely.''), <a href="https://www.ftc.gov/sites/default/files/documents/cases/2008/01/080122kovacic.pdf">https://www.ftc.gov/sites/default/files/documents/cases/2008/01/080122kovacic.pdf</a>.
\3\ Cf. Concurring and Dissenting Statement of Commissioner
Melissa Holyoak, Social Media and Video Streaming Services Staff
Report, Matter No. P205402, at 4 (Sept. 19, 2024), <a href="https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-social-media-6b.pdf">https://www.ftc.gov/system/files/ftc_gov/pdf/commissioner-holyoak-statement-social-media-6b.pdf</a>.
---------------------------------------------------------------------------
Some background first. Rytr provides AI products that generate
draft written content for 43 ``use cases,'' which are tailored written
outputs intended for specific purposes, like ``Email,'' ``Product
Description,'' ``Blogs,'' ``Articles,'' or ``Story Plot,'' among
[[Page 80567]]
others.\4\ Users can review and manually copy, paste, and modify the
draft content Rytr's products generate.\5\ Rytr gives all users access
to outputs from all 43 types of use cases, but certain paying
subscribers can generate unlimited output from all of them.\6\ As
alleged in the complaint, the ``Testimonial & Review'' feature, which
Rytr has stopped offering,\7\ generated draft consumer reviews based on
a user's inputs. Subscribers would provide descriptive keywords,
phrases, or titles, as well as choose the output's ``desired tone'' and
level of ``creativity,'' and which of dozens of languages to draft
in.\8\ Subscribers could opt to receive up to three ``variant'' drafts
at a time for comparison.\9\ The complaint assumes users mechanically
posted consumer reviews from Rytr's drafts without reviewing or
modifying them for accuracy. In other words, the complaint effectively
treats draft outputs as final reviews--reviews that users posted
without scrutiny.\10\ The complaint suggests Rytr did not monitor, or
provide feedback on, the content of automated drafts. The complaint
quotes several draft reviews and describes others. Finally, the
complaint alleges that 24 subscribers each generated over 10,000 draft
reviews; 114 subscribers each generated over 1,000 draft reviews; and
630 subscribers each generated over 100 draft reviews.\11\
---------------------------------------------------------------------------
\4\ Compl. ]] 2, 4.
\5\ See id. ] 4. The complaint does not allege that Rytr tracks
or knows whether or how user-directed draft content is ultimately
used.
\6\ See id. ]] 2, 5.
\7\ See Use Cases, Rytr, available at <a href="https://rytr.me/use-cases/">https://rytr.me/use-cases/</a>
(last visited Sept. 24, 2024).
\8\ Compl. ] 6; see also Languages, Rytr (``Rytr supports 30+
languages''), available at <a href="https://help.rytr.me/knowledge-base/languages">https://help.rytr.me/knowledge-base/languages</a> (last visited Sept. 24, 2024).
\9\ See id. ] 6.
\10\ The complaint's treatment of Rytr's product seems
implausible given Rytr's product design. If Rytr's intent was that
drafts be posted blindly, why would Rytr have given users up to
three ``variant'' drafts to consider and compare? See id.
\11\ See id. ] 13. The complaint does not allege whether there
were users that paid for Rytr's ``unlimited'' draft outputs across
all Rytr's use cases but who generated comparatively fewer draft
reviews than the volumes the complaint specifies. For example, it is
unclear whether certain users with access to unlimited outputs
generated high volumes of draft emails or blog posts, and also
drafted some reviews.
---------------------------------------------------------------------------
Critically, the complaint does not allege that users actually
posted any draft reviews.\12\ Since the Commission has no evidence that
a single draft review was posted, the complaint centers on alleging
speculative harms that may have come from subscribers with access to
unlimited output from across Rytr's use cases, which included draft
reviews.\13\ And the complaint asserts the review tool in particular
had ``no or de minimis reasonable, legitimate use,'' and concludes that
a function enabling unlimited reviews with limited user input, which
could generate ``detailed and genuine-sounding reviews,'' was likely to
``only . . . facilitate subscribers posting fake reviews . . . to
deceive consumers.'' \14\
---------------------------------------------------------------------------
\12\ In general, the complaint also does not address exactly how
much time passed as users generated drafts. Id. For example, for the
630 subscribers generating over 100 draft reviews, it is unclear
whether those reviews were spaced out over months or years.
\13\ Id. ]] 2, 5.
\14\ Id. ] 14.
---------------------------------------------------------------------------
Under the FTC Act, an act or practice is unfair if it ``causes or
is likely to cause substantial injury to consumers which is not
reasonably avoidable by consumers themselves and not outweighed by
countervailing benefits to consumers or to competition.'' \15\ In
relevant part, the complaint alleges that ``[Rytr] offered a service
intended to quickly generate unlimited content for consumer reviews'';
that it ``created false and deceptive written content for consumer
reviews''; and that any injury from Rytr's practices was ``not
outweighed by countervailing benefits to consumers or competition.''
\16\
---------------------------------------------------------------------------
\15\ 15 U.S.C. 45(n).
\16\ Compl. ] 17.
---------------------------------------------------------------------------
As a threshold matter, I am skeptical there is a likelihood of
substantial injury in this case, based on the facts pled, from Rytr's
offering unlimited drafts as part of a suite of over forty generative
AI offerings. For one thing, there is no concrete allegation that any
of the draft content generated in response to users' prompts was itself
false or inaccurate.\17\ And even if there were a question about the
drafts generated by Rytr, the complaint makes no allegation that any
drafts were ever posted online. To satisfy the substantial injury prong
of unfairness therefore, the complaint assumes, but does not allege
facts showing, that all users mechanically posted drafts without
modifying them to refine potential inaccuracies regarding the
underlying product and the user's experience with the product. If there
were in fact a likelihood of substantial injury, presumably the
complaint could allege at least one example where particular content
Rytr generated was untruthful and was actually posted to reach
consumers. Yet it does not. Unfairness requires proof--not
speculation--of harm.\18\
---------------------------------------------------------------------------
\17\ Instead, the complaint alleges that drafts were
``detailed'' and contain ``specific, often material details that
have no relation to a user's input,'' making it almost certain the
drafts would ``be false for the users who copy the generated content
and publish it online.'' Id. ] 8. The complaint also provides
several examples of users that generated high volumes of drafts,
including (apparently) a business, that seem likely to have been
fake if used. Id. ] 13. But again, the complaint alleges no facts
showing any such draft--let alone an unmodified draft--ever reached
consumers.
\18\ Cf. In re International Harvester Co., 104 F.T.C. 949,
1070, 1073 (1984) (unfairness statement) (``First of all, the injury
must be substantial. The Commission is not concerned with . . .
merely speculative harms.'').
---------------------------------------------------------------------------
Most significantly, though, by banning Rytr's user review service
the complaint fails to weigh the countervailing benefits Rytr's service
offers to consumers or competition. To start with the obvious, and as
the complaint itself describes, Rytr's tool generated ``reviews quickly
and with little user effort.'' \19\ It is generally a feature, not a
bug, when a new tool helps users save time and achieve their goals with
less work. Generative AI is no exception. In fact, much of the promise
of AI stems from its remarkable ability to provide such benefits to
consumers using AI tools. And as today's complaint accurately
acknowledges, consumers rely on reviews to make decisions in the
marketplace.\20\ If Rytr's tool helped users draft reviews about their
experiences that they would not have posted without the benefit of a
drafting aid, consumers seeing their reviews benefitted, too. There
would be similar benefit to competition.\21\
---------------------------------------------------------------------------
\19\ Compl. ] 7.
\20\ Id. ] 14 (explaining ``[c]onsumers rely on reviews for fair
and accurate information about products and services'').
\21\ It is undoubtedly true that ``[h]onest competitors who do
not post fake reviews can lose sales to businesses that do . . . .''
Id. But honest competitors, whose customers may have used this tool
to post reviews they might not have otherwise, would also benefit
from the additional reviews Rytr may have enabled.
---------------------------------------------------------------------------
The complaint fails to account for other potential benefits that
are less obvious. When it comes to AI, it is arguably a feature of the
service that draft reviews may not fit the user's experience in their
draft form. As with any draft document, what's not included in a final
version can still create value if the initial draft prompts a new line
of thought. For example, perhaps someone reviewing dog shampoo might
not have thought to mention a shampoo's effect on shedding absent the
prompt of an AI-generated initial draft.\22\ But on reading that draft,
perhaps the consumer recalls their pet's shedding has actually
increased, and they modify their review accordingly--helping to flesh
out their review and better inform interested readers of downsides of
the product. Part of generative AI's promise is its ability to suggest
new lines of thought that may
[[Page 80568]]
never have occurred to a user in the first place. But recognizing such
benefits here requires treating Rytr's product accurately--as a
drafting aid. The complaint also fails to grapple with how Rytr offered
unlimited outputs across a suite of over 40 products. Given generative
AI's manifold applications, there may be significant benefit to
consumers and competition when a company bundles its offerings and
their features so that users do not bump up against word restrictions
or character counts.\23\ The complaint today does not account for or
attempt to weigh such benefits. Instead, it baldly alleges there are
``no legitimate benefits'' from Rytr's service.\24\ That is mistaken
based on the facts pled, and a misapplication of our unfairness
authority.
---------------------------------------------------------------------------
\22\ Id. ] 10.
\23\ Even if a subscriber only chose to use the drafting tool
for reviews, there is no allegation that subscribers could buy
access only to unlimited drafting for reviews--undercutting the
complaint's suggestion that Rytr ``intended'' this particular
offering, on its own, ``to quickly generate unlimited content for
consumer reviews.'' See id. ] 17.
\24\ Id. ] 14.
---------------------------------------------------------------------------
Such observations about countervailing benefits lead to my concern
with the scope of today's order. Even if the Commission adequately pled
a law violation here, the Commission's order goes too far in its ban on
Rytr's providing any review or testimonial service. The complaint
alleges that Rytr's service was potentially misused by users to create
misleading reviews--not that the neutral service itself is a source of
harm. Banning products that have useful features but have the potential
to be misused is not consistent with the Commission's unfairness
authority. Nor is it consistent with a legal environment that promotes
innovation. AI is a developing industry. It has vast potential. We
should take care not to squelch it by suggesting that merely providing
draft content that could be used unlawfully is wrong.
Finally, I also share Commissioner Ferguson's views regarding the
complaint's ``means and instrumentalities'' claim.\25\ I write
separately to emphasize my concerns for imposing primary liability
under a means and instrumentalities claim, where there is no allegation
that Rytr itself made misrepresentations. The ``critical element'' for
primary liability ``is the existence of a representation, either by
statement or omission, made by the defendant . . . .'' \26\ The
complaint does not allege facts showing that the draft outputs were
misrepresentations, much less that such draft outputs were Rytr's
misrepresentations.\27\ Indeed, as the complaint alleges, a review can
also be deceptive if the user represents they experienced the product
when they never did.\28\ But such ``false'' reviews would also not be
misrepresentations made by Rytr.
---------------------------------------------------------------------------
\25\ Dissenting Statement of Commissioner Andrew N. Ferguson,
Joined by Commissioner Melissa Holyoak, In the Matter of Rytr LLC,
Matter No. 2323052 (Sept. 25, 2024).
\26\ Cf. Anixter v. Home-Stake Production Co., 77 F.3d 1215,
1225 (10th Cir. 1996); see also In re Shell Oil Co., 128 F.T.C. 749,
764 (1999) (majority statement) (``It is well settled law that the
originator is liable if it passes on a false or misleading
representation with knowledge or reason to expect that consumers may
possibly be deceived as a result.'') (citing Regina Corp. v. FTC,
322 F.2d 765 768 (3d Cir. 1963) (affirming liability under means and
instrumentalities theory where defendant distributed its own
misrepresentative price lists that were used, in turn, to deceive
consumers)); id. at 766 (Commissioner Swindle, dissenting) (``Means
and instrumentalities is a form of primary liability, and a
respondent is primarily liable only for its own misrepresentations
to consumers.'').
\27\ Users, not Rytr, chose any given review's tone (e.g.,
``critical'' or ``cautionary'' or ``convincing,'' etc.). Compl. ] 6.
Users, not Rytr, decided whether or not the draft review of a
particular product would be favorable or not. Id. And users, not
Rytr, decided whether to revise, or ultimately post, any given
draft. See id.
\28\ See id. ] 8.
---------------------------------------------------------------------------
In addition, the means and instrumentality claim cannot be
sustained on the theory that Rytr's AI service was somehow inherently
unfair or deceptive.\29\ Treating Rytr's neutral drafting tool as
inherently unfair or deceptive will have deleterious consequences for
AI products generally. Today's complaint suggests to all cutting-edge
technology developers that an otherwise neutral product used
inappropriately can lead to liability--even where, like here, the
developer neither deceived nor caused injury to a consumer.
---------------------------------------------------------------------------
\29\ Peerless Prods., Inc. v. FTC, 284 F.2d 825, 826 (7th Cir.
1960) (affirming Commission's findings that petitioners' punchboard
products were an unlawful ``means of and instrumentalities for
engaging in unfair acts'' where the punchboards were ``designed and
used primarily for the distribution of merchandise by lottery'').
---------------------------------------------------------------------------
We must protect consumers through robust enforcement. Indeed, the
Commission is at its best when it does so. But we must also think
carefully about the potential harms to consumers and innovation that
attend misguided enforcement.\30\ Today's misguided complaint and its
erroneous application of section 5 will likely undermine innovation in
the AI space. I therefore respectfully dissent.
---------------------------------------------------------------------------
\30\ Cf. Concurring and Dissenting Statement of Commissioner
Melissa Holyoak, Social Media and Video Streaming Services Staff
Report, supra note 3, at 18-19.
---------------------------------------------------------------------------
Dissenting Statement of Commissioner Andrew N. Ferguson Joined by
Commissioner Melissa Holyoak
The Commission today issues an administrative complaint and accepts
a proposed consent agreement with Rytr LLC (``Rytr'').\1\ Rytr has
created and markets a package of over 40 generative artificial
intelligence (``AI'') tools with a variety of uses, from writing essays
to creating poetry and music lyrics. One of these tools allowed users
to generate consumer reviews based on prompts provided by the user. For
having offered this tool, the Commission accuses Rytr of violating
section 5 of the Federal Trade Commission Act \2\ by furnishing its
users with the ``means and instrumentalities'' to deceive consumers.\3\
The Commission reasons that a business could use Rytr's tool to create
false or deceptive consumer reviews that the business could then pass
off as authentic reviews in violation of section 5. Rytr has agreed to
settle the case by promising not to offer similar functionality in the
future.
---------------------------------------------------------------------------
\1\ In re Rytr LLC, Complaint (``Complaint'') & Decision and
Order (``Order'').
\2\ 15 U.S.C. 45(a).
\3\ Complaint ] 15-16.
---------------------------------------------------------------------------
I dissent \4\ from the filing of the complaint and consent
agreement because I do not have reason to believe that Rytr violated
section 5, and because I do not believe filing is in the public
interest. The Commission's theory is that section 5 prohibits products
and services that could be used to facilitate deception or unfairness
because such products and services are the means and instrumentalities
of deception and unfairness. Treating as categorically illegal a
generative AI tool merely because of the possibility that someone might
use it for fraud is inconsistent with our precedents and common sense.
And it threatens to turn honest innovators into lawbreakers and risks
strangling a potentially revolutionary technology in its cradle.
---------------------------------------------------------------------------
\4\ In this statement, I discuss the Commission's charge of
deceptive conduct against Rytr. I also join Commissioner Holyoak's
dissent ably disposing of the Complaint's charge of unfair conduct.
Dissenting Statement of Commissioner Melissa Holyoak, In re Rytr,
LLC (Sept. 25, 2024).
---------------------------------------------------------------------------
I
Rytr is a generative AI toolkit designed to help users write and
edit text. Rytr markets its offering as including more than 40
different tools, which it calls ``use cases,'' that generate draft
emails, text messages, letters, advertisements, product descriptions,
blogs, articles, poems, song lyrics,
[[Page 80569]]
business pitches, and job descriptions.\5\ To generate content, the
user selects a use case and fills out a form asking for the parameters
of the text to be generated, such as the language, tone, creativity
level, and the basic idea governing the document (depending on the use
case, the business idea, the idea for a song, a description of the job
role, and so forth).\6\ Rytr's generative AI system processes the
information and produces a document in the number of variants requested
by the user.
---------------------------------------------------------------------------
\5\ Complaint ] 4; Use Cases, Rytr (last visited Sept. 11,
2024), <a href="https://rytr.me/use-cases/">https://rytr.me/use-cases/</a>.
\6\ The complaint contains a screenshot of the form for the
Testimonial & Review use case. Complaint ] 6. The Rytr website also
shows what the forms for the various use cases look like. Use Cases,
Rytr (last visited Sept. 11, 2024), <a href="https://rytr.me/use-cases/">https://rytr.me/use-cases/</a>
(click on a use case to see the form for that use case).
---------------------------------------------------------------------------
Rytr offers both free and paid versions of its writing tool. The
free version substantially limits the monthly character count.\7\ For
$9 per month (or $90 per year), a user can generate up to 100,000
characters.\8\ For $29 per month (or $290 per year), the user can
generate unlimited content.\9\ A user does not purchase subscriptions
for particular use cases. Rather, a single subscription gives the user
access to all of Rytr's use cases.\10\
---------------------------------------------------------------------------
\7\ Complaint ] 4.
\8\ Ibid.
\9\ Ibid.
\10\ Id. ] 5.
---------------------------------------------------------------------------
Until recently, one of Rytr's use cases was ``Testimonial &
Review.'' This use case permitted a user to generate product reviews by
entering a description, some keywords, or a product title.\11\ Nothing
prevented a user with a full subscription from using the tool to
generate unlimited reviews of multiple products, or of a single
product.
---------------------------------------------------------------------------
\11\ Id. ] 6.
---------------------------------------------------------------------------
The Commission alleges that this use case ``generates detailed
reviews that contain specific, often material details that have no
relation to the user's input''; that such reviews ``would almost
certainly be false''; and that, if such a review were ``publish[ed] . .
. online,'' it ``would deceive potential consumers deciding to purchase
the service or product described'' in violation of section 5.\12\ The
Complaint does not identify a single Rytr-generated review published
anywhere by anyone, much less a false review that violates section 5.
It nevertheless concludes that Rytr ``has furnished its users and
subscribers with the means to generate written content for consumer
reviews that is false and deceptive.'' \13\ ``[F]urnishing others with
the means and instrumentalities to engage in'' deception, the
Commission declares, is a section 5 violation.\14\
---------------------------------------------------------------------------
\12\ Id. ] 8.
\13\ Id. ] 15.
\14\ Id. ] 16.
---------------------------------------------------------------------------
II. A
Section 5 prohibits ``deceptive acts or practices.'' \15\ ``The
[Commission] must show three elements to prove a deceptive act or
practice in violation of Section 5(a)(1): [1] a representation,
omission, or practice, that [2] is likely to mislead consumers acting
reasonably under the circumstances, and [3], the representation,
omission, or practice is material.'' \16\ Although the Commission need
not show that the challenged representation or omission in fact
deceived a particular consumer,\17\ the Commission must show that the
defendant made a material misrepresentation or omission that was likely
to mislead consumers.\18\
---------------------------------------------------------------------------
\15\ 15 U.S.C. 45(a).
\16\ FTC v. Moses, 913 F.3d 297, 306 (2d Cir. 2019) (quotation
marks omitted); accord FTC v. Stefanchik, 559 F.3d 924, 928 (9th
Cir. 2009) (similar); POM Wonderful, LLC v. FTC, 777 F.3d 478, 490
(D.C. Cir. 2015) (similar); FTC v. Freecom Commc'ns, Inc., 401 F.3d
1192, 1203 (10th Cir. 2005) (similar).
\17\ FTC v. <a href="http://Cyberspace.com">Cyberspace.com</a>, 453 F.3d 1196, 1201 (9th Cir. 2006).
\18\ FTC v. Tashman, 318 F.3d 1273, 1277 (11th Cir. 2003).
---------------------------------------------------------------------------
The Commission does not allege that Rytr made a misleading
statement or omission of any kind, much less one that was material or
likely to mislead consumers. The Commission instead pleads that Rytr
furnished the ``means and instrumentalities'' by which someone else
could make false statements in violation of section 5.
Means-and-instrumentalities liability arises from a century-old
case involving not ``unfair or deceptive acts or practices,'' but
``unfair methods of competition.'' In FTC v. Winsted Hosiery Co., the
Supreme Court considered an unfair-method-of-competition claim against
a clothing company that falsely labeled its clothing as being made of
wool.\19\ The company defended on the ground that the retailers to whom
it sold its clothing were fully aware that the labels were false, and
therefore the company did not deceive anyone.\20\ The Court rejected
this defense, holding that the clothier was a ``wrongdoer'' because it
``furnishe[d]'' retailers ``with the means of consummating a fraud''
against consumers, who were not aware that the labels were false.\21\
Courts and the Commission have since relied on Winsted Hosiery to hold
that a person ``who puts into the hands of others the means by which
such others may deceive the public is equally responsible for the
resulting deception.'' \22\ This theory of liability has come to be
known as ``means-and-instrumentalities'' liability, and ``is intended
to apply in cases . . . where the originator of the unlawful material
is not in privity with consumers.'' \23\
---------------------------------------------------------------------------
\19\ 258 U.S. 483, 490-91 (1922).
\20\ Id. at 492-93.
\21\ Ibid.
\22\ In re Litton Indus., Inc., 97 F.T.C. 1, 46-47 (1981); see
also, e.g., FTC v. Magui Publishers, Inc., 9 F.3d 1551, 1993 WL
430102, at *4 (9th Cir. 1993) (unpublished) (``It is well
established that one who puts into the hands of others the means by
which such others may deceive the public is equally as responsible
for the resulting deception.''); C. Howard Hunt Pen Co. v. FTC, 197
F.2d 273, 281 (3d Cir. 1952) (``One who places in the hands of
another a means of consummating a fraud or competing unfairly in
violation of the Federal Trade Commission Act is himself guilty of a
violation of the Act.''); Waltham Watch Co. v. FTC, 318 F.2d 28, 32
(7th Cir. 1963) (``Those who put into the hands of others the means
by which they may mislead the public, are themselves guilty of a
violation of Section 5 of the Federal Trade Commission Act.'').
\23\ In re Shell Oil Co., 128 F.T.C. 749, 764 (1999).
---------------------------------------------------------------------------
Means-and-instrumentalities liability has traditionally been
confined to two types of cases. The first involves a defendant who
supplies someone other than a consumer--ordinarily a retailer--with a
product or service that is unlawful because it is inherently deceptive,
or because it has no purpose apart from facilitating a section 5
violation. The recipient of that product or service then passes it on
to consumers in violation of section 5. Winsted Hosiery was such a
case. The Commission relied on this theory for many decades to pursue
makers of push cards and punch boards custom-made for retailers to use
in illegal lottery marketing schemes.\24\ It has also relied on this
theory to pursue suppliers of mislabeled art, which retailers then sold
to deceived consumers.\25\
---------------------------------------------------------------------------
\24\ See, e.g., Gellman v. FTC, 290 F.2d 666, 667-68 (8th Cir.
1961) (collecting cases); Peerless Prods., Inc. v. FTC, 284 F.2d
825, 826 (7th Cir. 1960); James v. FTC, 253 F.2d 78, 80 (7th Cir.
1958); Globe Cardboard Novelty Co. v. FTC, 192 F.2d 444, 446 (3d
Cir. 1951); Chas. A. Brewer & Sons v. FTC, 158 F.2d 74, 77 (6th Cir.
1946); FTC v. F.A. Martoccio Co., 87 F.2d 561, 564 (8th Cir. 1937).
\25\ See, e.g., Magpui, 1993 WL 430102, at *4; Int'l Art Co. v.
FTC, 109 F.2d 393, 397 (7th Cir. 1940).
---------------------------------------------------------------------------
The second type of means-and-instrumentalities case involves
suppliers of misleading marketing materials that someone down the
supply chain uses to deceive consumers. In these cases, the defendant
makes false or misleading statements to someone further down the supply
chain, who then repeats the misstatements to deceive consumers.\26\ If
the repeated
[[Page 80570]]
statement does not satisfy the three-part test for deception under
section 5, however, it cannot give rise to means-or-instrumentalities
liability.\27\ The classic example of this case involves deceptive
marketing materials for multilevel-marketing businesses and ``pyramid''
schemes. The participants at the top of the pyramid do not interact
with consumers; they instead convey false statements to others further
down the pyramid who in turn use those materials to deceive consumers.
The Commission has used the means-and-instrumentalities theory against
the orchestrators of deception who sit at the top of the pyramid.\28\
---------------------------------------------------------------------------
\26\ See, e.g., Regina Corp. v. FTC, 322 F.2d 765, 768 (3d Cir.
1963) (supplier's conveying of a deceptive list price to retailers,
which was repeated to consumers).
\27\ FTC v. Innovative Designs, Inc., No. 20-3379, 2012 WL
3086188, at *4 n. 11 (3d Cir. July 22, 2021).
\28\ See, e.g., FTC v. Noland, 672 F. Supp. 3d 721, 786 (D.
Ariz. 2023); FTC v. Fin. Educ. Servs., Inc., No. 2:22-CV-11120, 2022
WL 19333298, at *1 (E.D. Mich. May 24, 2022); FTC v. Neora LLC, 552
F. Supp. 3d 628 (N.D. Tex. 2021); FTC v. Vemma Nutrition Co., No.
15-cv-1578, 2015 WL 11118111, at *7 (D. Ariz. Sept. 18, 2015); FTC
v. <a href="http://Skybiz.com">Skybiz.com</a>, Inc., No. 01-CV-396-K(E), 2001 WL 1673645, at *1
(N.D. Okla. Aug. 31, 2001); FTC v. Five-Star Auto Club, Inc., 97 F.
Supp. 2d 502, 530-31 (S.D.N.Y. 2000).
---------------------------------------------------------------------------
This categorization seems straightforward at first blush, but the
means-and-instrumentalities doctrine becomes less coherent the closer
one looks. On the one hand, we have described `` `means and
instrumentalities' liability [as] a form of direct liability,'' \29\
that is, as a way of holding someone ``directly liable for violating''
\30\ Section 5 ``distinct from `aiding and abetting' liability and
`assisting and facilitating' liability, both of which are secondary
forms of liability.'' \31\ That appears to be true when the Commission
uses this theory against the orchestrator of a pyramid scheme, who
makes misrepresentations to someone other than a consumer but which
misrepresentations are repeated to consumers by people further down the
pyramid.\32\ When applying means-and-instrumentalities liability
against defendants who supplied the component parts of someone else's
section 5 violation, however, courts have described the theory as a
species of aiding-and-abetting liability.\33\ We have also told
Congress that our means-and-instrumentalities theory is an
``alternative theor[y]'' to aiding-and-abetting liability by which we
can ``reach secondary actors.'' \34\ Means-and-instrumentalities
liability therefore sometimes functions as a form of direct liability
(when deployed against the orchestrator of a pyramid scheme, for
example) and sometimes as a form of aiding-and-abetting liability (when
deployed against the makers of punch boards and push cards, for
example).
---------------------------------------------------------------------------
\29\ FTC, SNPRM: Trade Regulation Rule on Impersonation of
Government and Businesses (``SNPRM''), 89 FR 15072, 15077 n.94 (Mar.
1, 2024).
\30\ FTC v. Magui Publishers, Inc., No. Civ. 89-3818, 1991 WL
90895, at *14 (C.D. Cal. Mar. 28, 1991), aff'd, 9 F.3d 1551 (9th
Cir. 1993).
\31\ SNPRM, 89 FR 15077 n.94.
\32\ See Shell Oil, 128 F.T.C. at 766 (dissenting statement of
Commissioner Swindle) (``Means and instrumentalities is a form of
primary liability, and a respondent is primarily liable only for its
own misrepresentations to consumers.'').
\33\ See, e.g., Chas. A Brewer & Sons, 158 F.2d at 77
(describing ``furnishing the means of consummating a fraud'' as
``aiding and abetting'' another's ``unfair or deceptive acts or
practices''); Gay Games, Inc. v. FTC, 204 F.2d 197, 199 (10th Cir.
1953) (similar); Consol. Mfg. Co. v. FTC, 199 F.2d 417, 418 (4th
Cir. 1952) (per curiam) (similar); see also Deer v. FTC, 152 F.2d
65, 66 (2d Cir. 1945) (``[I]t was not necessary to prove that the
petitioners actually participated in the operation of the bingo game
or the club plan conducted by their customers; it is enough that
they aided and abetted in such'' games by furnishing the
paraphernalia for the game.).
\34\ Federal Trade Commission Reauthorization: Hearing Before
the S. Comm. on Com., Sci., and Transp., S. Hrg. 110-1148, p. 21
n.56 (2008) (Prepared Statement of the Federal Trade Commission).
---------------------------------------------------------------------------
The complaint against Rytr falls into neither category. The
Commission does not accuse Rytr of making any statements, much less
false statements. Nor is Rytr's tool necessarily deceptive like
mislabeled art, or useful only in facilitating someone else's section 5
violation like lottery punch boards. Rytr's tool has both lawful and
unlawful potential uses. A consumer could use it to draft an honest and
accurate review. Or a business could use it to write a false review.
B. 1
The Commission's complaint is a dramatic extension of means-and-
instrumentalities liability. The Commission treats Rytr's sale of a
product with lawful and unlawful potential uses as a categorical
section 5 violation because someone could use it to write a statement
that could violate section 5. But that is true of an almost unlimited
number of products and services: pencils, paper, printers, computers,
smartphones, word processors, typewriters, posterboard, televisions,
billboards, online advertising space, professional printing services,
etc. On the Commission's theory, the makers and suppliers of these
products and services are furnishing the means or instrumentalities to
deceive consumers merely because someone might put them to unlawful
use.
This theory is incorrect. Section 5 does not categorically prohibit
a product or service merely because someone might use it to deceive
someone else. Interpreting section 5 to prohibit products and services
with conceivable illegal uses would prohibit an infinite variety of
innocent and productive conduct. Congress cannot have intended to
capture such conduct in the phrase ``deceptive acts and practices.''
Not only is the Commission's theory a departure from section 5
precedents, but it is also inconsistent with how other areas of the law
deal with the same issue. In Sony Corp. of America v. Universal City
Studios, Inc., for example, the Supreme Court considered whether a
product capable of facilitating en masse copyright infringement--
Betamax video recorders capable both of lawfully playing Betamax tapes
and of unlawfully recording copyrighted television broadcasts--violated
the copyright laws.\35\ The Court concluded that so long as a product
is ``merely . . . capable of substantial noninfringing uses,'' it did
not violate the copyright laws even if it is also capable of committing
countless acts of infringement.\36\ Similarly, patent law does not
treat as infringement the sale of an unpatented part of a patented
machine that could be used to infringe the patent, so long as the part
is capable of some noninfringing uses.\37\
---------------------------------------------------------------------------
\35\ 464 U.S. 417, 419-20, 421-23 (1984).
\36\ Id. at 442.
\37\ See, e.g., Dawson Chem. Co. v. Rohm & Hass Co., 448 U.S.
176, 198 (1980) (holding that the patent laws do not forbid the sale
of ``unpatented articles that were essential to . . . patented
inventions'' unless those unpatented articles ``were unsuited for
any commercial noninfringing use''); Henry v. A.B. Dick Co., 224
U.S. 1, 48 (1912) (``[A] sale of an article which though adapted to
an infringing use is also adapted to other and lawful uses, is not
enough to make the seller a contributory infringer. Such a rule
would block the wheels of commerce.'')
---------------------------------------------------------------------------
Aiding-and-abetting liability, which bears many similarities to
means-and-instrumentalities liability,\38\ also does not punish conduct
merely because it facilitated the commission of a tort or crime.
Liability for aiding and abetting under Federal criminal law requires
``that the accused ha[d] the specific intent to facilitate the
commission of a crime by another'' as well as ``the requisite intent of
the underlying substantive offense.'' \39\ And in tort law, one is
liable for the torts of another ``if he knows that the other's conduct
constitutes a breach of duty and gives substantial assistance or
encouragement to the other.'' \40\
---------------------------------------------------------------------------
\38\ See supra n. 29-34 and accompanying text.
\39\ U.S. Department of Justice, Criminal Resource Manual, Sec.
2474. Elements of Aiding and Abetting (synthesizing Federal
appellate precedent).
\40\ Restatement (Second) of Torts Sec. 876(b) (1979) (emphasis
added).
---------------------------------------------------------------------------
[[Page 80571]]
2
The Commission tries to diminish the grandiosity of its theory by
alleging that Rytr's tool ``has no or de minimis legitimate use.'' \41\
If this were true, then I might agree with the Commission's decision to
file this complaint. Courts have for decades interpreted section 5 to
prohibit the sale of products with no reasonable uses other than
facilitating an unfair or deceptive act or practice.\42\ But the
Commission's conclusory description of the Rytr tool's plausible uses
is pure ipse dixit. The complaint contains no factual allegations
lending plausibility to its conclusion that the tool has no, or only de
minimis, legitimate uses.\43\ Nor I have seen any evidence giving me
reason to believe that the allegation is true.
---------------------------------------------------------------------------
\41\ Complaint ] 14.
\42\ See supra n. 24-25 and accompanying text (discussing courts
of appeals precedents sustaining Commission orders prohibiting the
sale of push cards and punch boards).
\43\ Ashcroft v. Iqbal, 556 U.S. 662, 678 (2009) (``Threadbare
recitals of the elements of a cause of action, supported by mere
conclusory statements, do not suffice'' for a complaint to survive a
motion to dismiss.).
---------------------------------------------------------------------------
Indeed, the complaint's conclusion is entirely implausible. For one
thing, if the Rytr tool's exclusive use were to generate false consumer
reviews in violation of section 5, one would expect the complaint to
contain allegations that someone used it to violate section 5, at least
once. But the Commission does not allege a single example of a Rytr-
generated review being used to deceive consumers in violation of
section 5, nor am I aware of any.
The Rytr tool's legitimate utility to consumers is obvious: to
assist them in writing reviews. Writing a succinct and thoughtful
review can be difficult and time-consuming,\44\ and a tool that
produces a well-written first draft of a review based on some keyword
inputs can make the task much more accessible.
---------------------------------------------------------------------------
\44\ See Blaise Pascal, Letter XVI, Lettres Provinciales (1657)
(``I would have written a shorter letter, but I did not have the
time.'').
---------------------------------------------------------------------------
The Commission describes the Rytr tool's only use as ``generating
written content for a review'' that a user would then ``manually select
and copy . . . to post reviews elsewhere online.'' \45\ But consumers
do not have to use generative AI as a replacement for their own
thoughts and ideas. Consumers can use AI-generated first drafts of
documents in much the same way they would use a human-generated first
draft--as a starting point from which the user can work to convey
accurately and clearly the idea in the user's mind. A consumer would
not violate section 5 by using a generative AI tool to write a first
draft of a review, even if that first draft contained inaccuracies that
the user then removed.
---------------------------------------------------------------------------
\45\ Complaint ] 6; see also id. ] 8 (``Respondent's service
generates reviews that would almost certainly be false for the users
who copy the generated content and publish it online.'').
---------------------------------------------------------------------------
I do not doubt that some people use generative AI tools to
accomplish fraud. Almost every technology since the first time a human
being sharpened a stick can be put to some illegal use. But that does
not mean that those tools are the means and instrumentalities to
deceive consumers. Section 5 does not prohibit the sale of any product
that someone could use to violate section 5.
C
The question, then, is whether the Commission may ever treat a
product or service with lawful and unlawful potential uses as the means
and instrumentalities to violate section 5. The law is clear that it
may, but only if the provider of the product or service knows, or has
reason to know, that the person to whom the product or service was
supplied will use it to violate section 5. A knowledge requirement
avoids treating innocent and productive conduct as illegal merely
because of the subsequent acts of independent third parties.
Courts have required knowledge in similar means-and-
instrumentalities cases for decades. In Waltham Watch Co. v. FTC, for
example, a clockmaker granted a license to use the clockmaker's famous
trademark in the sale of clocks.\46\ The licensee then used the
trademark to deceive consumers and other dealers into believing that
clocks had been manufactured by the clockmaker.\47\ The clockmaker was
liable for the deception, the court of appeals reasoned, because the
clockmaker in fact knew that the licensee was using the license to
commit fraud and took no action to prevent it.\48\ That knowledge
transformed an otherwise neutral license agreement into the ``means and
instrumentalities by which many people had been hoodwinked, defrauded,
and misled.'' \49\ Similarly, in one of the Commission's most important
statements of the scope of means-and-instrumentalities liability, we
explained that ``[i]t is well settled law'' under the means-and-
instrumentalities doctrine that ``the originator'' of a false or
misleading representation ``is liable if it passes on a false or
misleading representation with knowledge or reason to expect that
consumers may possibly be deceived as a result.'' \50\
---------------------------------------------------------------------------
\46\ 318 F.2d 28, 30 (7th Cir. 1963).
\47\ Ibid.
\48\ Ibid.
\49\ Ibid.
\50\ Shell Oil, 128 F.T.C. at 764.
---------------------------------------------------------------------------
The Commission recently acknowledged that section 5 requires proof
of knowledge before treating products and services with lawful and
unlawful potential uses as the means and instrumentalities to violate
section 5. Earlier this year, the Commission promulgated the Trade
Regulation Rule on Impersonation of Government and Businesses
(Impersonation Rule).\51\ The rule treats the impersonation of a
government official or business as an unfair or deceptive act or
practice.\52\ Our Notice of Proposed Rulemaking (NPRM) for the
Impersonation Rule proposed treating the provision of ``the means and
instrumentalities for'' impersonation as a section 5 violation.\53\ But
we received a host of comments warning that imposing means-and-
instrumentalities liability without a scienter requirement would
``impos[e] strict liability on unwitting third-party providers of
services or products.'' \54\ We therefore removed the mean-and-
instrumentalities provision from the Impersonation Rule and issued a
supplemental NPRM on the same topic.\55\ The supplemental NPRM proposes
treating the provision of ``goods or services'' as a section 5
violation only if ``a party knew or had reason to know that the goods
or services they provided will be used for the purpose of
impersonations.\56\
---------------------------------------------------------------------------
\51\ 89 FR 15017 (Mar. 1, 2024) (to be codified at 16 CFR part
461).
\52\ 16 CFR 461.2, 461.3.
\53\ 87 FR 62741, 62751 (Oct. 17, 2022).
\54\ 89 FR at 15022.
\55\ 89 FR at 15023.
\56\ SNPRM, 89 FR 15072, 15077 (Mar. 1, 2024); see also id. at
15083 (text of proposed rule).
---------------------------------------------------------------------------
Section 5 also requires proof of knowledge of third-party behavior
in other, similar contexts. For example, a defendant is liable for the
deceptive acts of its third-party affiliates only if the defendant has
actual knowledge of the affiliates' ongoing deception and ``either
directly participates in that deception, or has the authority to
control'' it and ``allows the deception to proceed.'' \57\ And section
5 imposes liability on an individual officer for the violations of a
corporate entity only if ``the individual had `some knowledge of the
practices' and . . . either `participated directly in the practice or
acts or had the authority to control them.' '' \58\ These knowledge
[[Page 80572]]
requirements implement a common sense principle: section 5 does not
hold people liable for innocent conduct that may have unwittingly
facilitated someone else's violation.
---------------------------------------------------------------------------
\57\ FTC v. LeadClick Media, LLC, 838 F.3d 158, 170 (2d Cir.
2016).
\58\ FTC v. On Point Capital Partners LLC, 17 F.4th 1066, 1083
(11th Cir. 2021) (quoting FTC v. Gem Merch. Corp., 87 F.3d 466, 470
(11th Cir. 1996)); FTC v. Moses, 913 F.3d 297, 306-07 (2d Cir.
2019); FTC v. Com. Planet, Inc., 815 F.3d 593, 600 (9th Cir. 2016)
(similar); FTC v. Freecom Commc'ns, Inc., 401 F.3d 1192, 1203-04
(10th Cir. 2005) (similar); FTC v. Amy Travel Serv., Inc., 875 F.2d
564, 573 (7th Cir. 1989) (similar).
---------------------------------------------------------------------------
Other areas of the law abide by the same common-sense principle. In
Metro-Goldwyn-Mayer Studios Inc. v. Grokster, for example, the Supreme
Court again confronted the question of whether a product with both
infringing and noninfringing uses violated the copyright laws.\59\ In
that case, the product was peer-to-peer file sharing software that was
commonly used to share copyrighted music and films without
authorization.\60\ Although the copyright laws do not prohibit a
product ``capable of commercially significant noninfringing uses'' even
if it were also capable of substantial infringement,\61\ the makers of
the peer-to-peer filesharing software distributed their product with
the intention of promoting infringement.\62\ Imposing copyright
liability on a party who distributed a product with the intention of
facilitating infringement was consistent with ``principles recognized
in every part of the law.'' \63\
---------------------------------------------------------------------------
\59\ 545 U.S. 913 (2005).
\60\ Id. at 919-20.
\61\ Id. at 931-32.
\62\ Id. at 934-35.
\63\ Id. at 935 (quoting Kalem Co. v. Harper Bros., 222 U.S. 55,
63 (1911)).
---------------------------------------------------------------------------
The point here is not to identify exhaustively the circumstances in
which the provision of a product or service with lawful and unlawful
potential uses may violate section 5. I instead argue only that, at the
very least, precedent and common-sense ``principles recognized in every
part of the law'' require that the government must show that a
defendant knew that he was participating in someone else's unfair or
deceptive act or practice when he provided that product or service.
III
I dissent from the filing of this complaint for an additional
reason. We may file an administrative action alleging a section 5
violation only if such an action ``would be to the interest of the
public.'' \64\ I do not believe this action is in the public interest
for two reasons.
---------------------------------------------------------------------------
\64\ 15 U.S.C. 45(b).
---------------------------------------------------------------------------
First, the Commission's aggressive move into AI regulation is
premature. AI is the subject of heated rhetoric. Doomsayers warn that
AI will take our jobs, hopelessly blur the distinction between fact and
fiction, and maybe even threaten the survival of human civilization. AI
companies do not forcefully resist all these claims, given that
predictions about the incredible potential for AI may be useful as
these companies compete for investment dollars and engineering talent.
But the Commission should not succumb to the panic or hype. Generative
AI technology is impressive, but it is also nascent. Neither its
naysayers nor its cheerleaders really understand its potential, or
whether it represents substantial progress toward ``artificial general
intelligence'' (AGI)--machine intelligence matching both the breadth
and power of the human mind, the holy grail of AI research.\65\ That
ignorance is not a reason to plunge headlong with aggressive
regulation. It is a reason to stay our hand.
---------------------------------------------------------------------------
\65\ Concurring and Dissenting Statement of Commissioner Andrew
N. Ferguson, A Look Behind the Screens: Examining the Data Practices
of Social Media and Video Streaming Services, at 11 n.44 (Sept. 19,
2024).
---------------------------------------------------------------------------
As our country has always done, we should give this industry the
space to realize its full potential--whatever that turns out to be.
America is the greatest commercial power in the history of the world in
no small part because of its tolerant attitude toward innovation and
new industry. There has never been a better place in the world to have
a new idea than the United States. We should go to great lengths to
ensure that remains the case.
When people use generative AI technology to lie, cheat, and steal,
the law should punish them no differently than if they use quill and
parchment.\66\ But Congress has not given us the power to regulate AI.
It has tasked us with enforcing the prohibition against unfair or
deceptive acts and practices. If our enforcement incidentally captures
some AI-generated conduct, so be it.\67\ But we should not bend the law
to get at AI. And we certainly should not chill innovation by
threatening to hold AI companies liable for whatever illegal use some
clever fraudster might find for their technology.
---------------------------------------------------------------------------
\66\ Id. at 10-11.
\67\ I support, for example, the complaint and settlement that
we announce today against DoNotPay for deceiving consumers about the
capabilities of its generative AI service. Concurring Statement of
Commissioner Andrew N. Ferguson, In the Matter of DoNotPay, Inc.
(Sept. 25, 2024).
---------------------------------------------------------------------------
Second, the complaint implicates important First Amendment
interests. The First Amendment constrains the government's authority to
regulate the inputs of speech.\68\ The Commission today holds a company
liable under section 5 for a product that helps people speak, quite
literally. The theory on which the complaint rests would permit the
Commission to proscribe Microsoft Word merely because someone may use
it to create a fake review, or Adobe Photoshop merely because someone
used it to create a false celebrity endorsement. The danger this theory
poses to free speech is obvious. Yet because the technology in question
is new and unfamiliar, I fear we are giving short shrift to common
sense and to fundamental constitutional values.
---------------------------------------------------------------------------
\68\ See, e.g., Buckley v. Valeo, 424 U.S. 1, 16, 19-20 & n.18,
44-45 (1976) (per curiam) (striking down Federal limitations on
political expenditures on the ground that such expenditures are a
necessary ingredient to the sort of mass political communication
protected by the Speech Clause); McConnell v. FEC, 540 U.S. 93, 251
(2003) (Scalia, J., concurring in part, concurring in the judgment
in part, and dissenting in part) (``To a government bent on
suppressing speech, this mode of organization presents
opportunities: Control any cog in the machine, and you can halt the
whole apparatus.''). See also Minneapolis Star & Tribune Co. v.
Minn. Comm'r of Revenue, 460 U.S. 575, 591-93 (1983) (striking down
a tax on paper and ink as an unconstitutional restriction of the
freedom of speech and of the press); Grosjean v. Am. Press Co., 297
U.S. 233, 250-51 (1936) (striking down statute taxing the sale of
advertisements in publications with a weekly circulation greater
than 20,000 copies).
---------------------------------------------------------------------------
I respectfully dissent.
[FR Doc. 2024-22767 Filed 10-2-24; 8:45 am]
BILLING CODE 6750-01-P
</pre></body>
</html>Indexed from Federal Register on October 3, 2024.
This is legal information, not legal advice. Laws vary by jurisdiction and change frequently. Always verify current law with official sources and consult a licensed attorney in your jurisdiction for advice on your specific situation.