The American Innovation and Choice Online
December 2, 2022
Act
Jay B. Sykes
The competitive practices of large technology companies have attracted considerable
Legislative Attorney
congressional attention in recent years. In October 2020, a House subcommittee concluded a
16-month investigation into the market power of four of the largest platform operators:
Facebook, Google, Amazon, and Apple. The inquiry culminated in a 450-page report
recommending a range of measures to address the allegedly anticompetitive conduct of these
“Big Tech” firms.
While the subcommittee’s report prompted a range of proposals, attention has turned to bills targeting discriminatory conduct
by the tech giants. The relevant legislation—the American Innovation and Choice Online Act (AICOA)—would prohibit Big
Tech platforms from favoring their own products and services in various ways.
There are different versions of the legislation. During the first session of the 117th Congress, the House Judiciary Committee
ordered a version of the AICOA to be reported to the full House. Bills embodying the legislation are also under active Senate
consideration in the second session. In January 2022, the Senate Judiciary Committee approved S. 2992, which it reported to
the full Senate in March. Senator Amy Klobuchar—S. 2992’s sponsor—later released a different version of the AICOA on
May 25, 2022. The May 25 draft indicates that Senator Klobuchar intends to introduce it as an amendment in the nature of a
substitute, if the Senate takes up the bill. This report provides an overview of the versions of the AICOA pending in the
Senate, while noting certain key differences between versions of the bill pending in the House and Senate.
S. 2992 would apply special rules to “covered platforms.” Under the May 25 draft, covered platforms would include online
platforms that exceed certain thresholds for U.S.-based active users; exceed certain thresholds for annual sales, market
capitalization, or worldwide active users; and occupy positions as “critical trading partners.”
The legislation would prohibit operators of covered platforms from engaging in 10 categories of conduct. Three of the
offenses would require regulators to establish that a platform operator’s conduct resulted in material harm to competition. In
particular, the legislation would bar operators of covered platforms from:
Preferencing their own products or services over those of other business users of their platforms in a
manner that would “materially harm competition”;
Limiting the ability of business users to compete with the operators’ own offerings in a manner that would
“materially harm competition”; and
Discriminating in the application of their terms of service among similarly situated business users in a
manner that would “materially harm competition.”
The remaining seven offenses involve a variety of other issues, including platform interoperability, use of user data, and
restrictions on the uninstallation of software applications. These offenses would not require regulators to prove competitive
harm, but defendants would be allowed to rebut a
prima facie case by establishing an
absence of such harm.
The legislation would also offer defendants several other affirmative defenses involving user privacy, data security, and
platform functionality. Enforcement authority would rest with the Department of Justice, the Federal Trade Commission, and
state attorneys general.
Congressional Research Service
link to page 5 link to page 9 link to page 10 link to page 10 link to page 11 link to page 12 link to page 16 link to page 17 link to page 19 link to page 20 link to page 22 link to page 22 link to page 22 link to page 23 link to page 24 link to page 25 link to page 26
The American Innovation and Choice Online Act (S. 2992)
Contents
Covered Platforms........................................................................................................... 2
Unlawful Conduct ........................................................................................................... 6
Sections 3(a)(1)-(3): Self-Preferencing, Limitations on Business Users, and
Discrimination That Would “Material y Harm Competition” .......................................... 7
Self-Preferencing ................................................................................................. 8
Material Harm to Competition................................................................................ 9
Content Moderation ............................................................................................ 13
Sections 3(a)(4) and 3(a)(7): Interoperability and Access to Data .................................... 14
Section 3(a)(5): Tying .............................................................................................. 16
Section 3(a)(6): Use of Data ...................................................................................... 17
Section 3(a)(8): App Preinstal ation and Steering .......................................................... 19
Section 3(a)(9): Self-Preferencing Involving Platform Interfaces..................................... 19
Section 3(a)(10): Retaliation for Reports to Law Enforcement ........................................ 19
Affirmative Defenses ..................................................................................................... 20
Enforcement................................................................................................................. 21
Conclusion................................................................................................................... 22
Contacts
Author Information ....................................................................................................... 23
Congressional Research Service
The American Innovation and Choice Online Act (S. 2992)
he competitive practices of large technology platforms have attracted considerable
congressional attention in recent years. In October 2020, a House subcommittee concluded
T a 16-month investigation into the market power of four of the largest platform operators:
Facebook, Google, Amazon, and Apple. The inquiry culminated in a 450-page report
recommending a range of measures to address the al egedly anticompetitive conduct of these
“Big Tech” firms.1
The subcommittee’s report prompted a flurry of legislative activity.2 The 117th Congress has
featured bil s that would impose vertical separation requirements,3 acquisition restrictions,4
interoperability and data-portability mandates,5 and a specialist regulator6 on Big Tech.
While these proposals cover a diverse range of topics, attention has turned to bil s targeting
discriminatory conduct by the tech giants.7 The relevant legislation—the American Innovation
and Choice Online Act (AICOA)—would prohibit Big Tech platforms from favoring their own
products and services in various ways.8 The prohibitions would move significantly beyond
existing antitrust doctrine and could have important ramifications for the shape of the digital
economy.9
There are different versions of the legislation. During the first session of the 117th Congress, the
House Judiciary Committee ordered a version of the AICOA—H.R. 3816—to be reported to the
full House.10 Bil s embodying the legislation are also under active consideration in the Senate in
the second session. In January 2022, the Senate Judiciary Committee approved S. 2992, which it
reported to the full Senate in March.11 Senator Amy Klobuchar—S. 2992’s sponsor—later
1 INVESTIGATION OF COMPETITION IN DIGITAL MARKETS, MAJORITY STAFF REPORT AND RECOMMENDATIONS, SUBCOMM.
ON ANTITRUST, COM. AND ADMIN. L. OF THE H. COMM. ON THE JUDICIARY, 116TH CONG. (2020). This report list s t he Big
T ech firms in the same order as the subcommittee’s report. Since the publication of the subcommittee’s report,
Facebook has changed its name to Meta Platforms, Inc.
2
See CRS Report R46875,
The Big Tech Antitrust Bills, by Jay B. Sykes.
3 H.R. 3825, 117th Cong. (2021).
4 S. 3197, 117th Cong. (2021); H.R. 3826, 117th Cong. (2021).
5 H.R. 3849, 117th Cong. (2021).
6 S. 4201, 117th Cong. (2022); H.R. 7858, 117th Cong. (2022).
7
See, e.g., Lauren Feiner,
Lawmakers Are Racing to Pass Tech Antitrust Reforms Before Midterms, CNBC (June 4,
2022), https://www.cnbc.com/2022/06/04/lawmakers-racing-to-pass-tech-antitrust-tech-reforms-before-midterms.html.
While “discrimination” can be a value-laden term, this report uses that language in a purely descriptive sense to refer to
conduct by a vertically integrated firm that preferences the firm’s own offerings over those of its rivals. T hat conduct is
“discriminatory” in this sense does not necessarily entail any conclusions about its competitive effects.
See Erik
Hovenkamp,
The Antitrust Duty to Deal in the Age of Big Tech , 131 YALE L.J. 1483, 1544 n.290 (2022) (collecting
academic literature debating the circumstances in which platform discrimin ation is anticompetitive);
cf. HERBERT
HOVENKAMP, FEDERAL ANTITRUST POLICY: THE LAW OF COMPETITION AND ITS PRACTICE 621-28 (2011) (discussing the
varied output effects of different types of price discrimination).
8
See Brendan Bordelon & Josh Sisco,
Schumer’s Office Says He Plans to Hold Vote on Tech Antitrust Bill, POLITICO
(Aug. 4, 2022), https://www.politico.com/news/2022/08/04/schumer-tech-antitrust -bill-00049890.
9
See, e.g., Letter from Fiona M. Scott Morton, et al., to Sen. Amy Klobuchar & Sen. Charles Grassley 5 (July 7, 2022),
https://som.yale.edu/sites/default/files/2022-07/AICOA-Final-revised.pdf (supporting the AICOA on the grounds that it
“significantly strengthens” the antitrust laws vis-à-vis Big T ech platforms); Ryan Bourne & Brad Subramaniam,
The
“Big Tech” Self-Preferencing Delusion, CATO INST. 2 (Feb. 24, 2022), https://www.cato.org/sites/cato.org/files/2022-
02/briefing-paper-136.pdf (opposing the AICOA and arguing that it would “ fundamentally alter business conduct on
covered platforms”).
10 H.R. 3816, 117th Cong. (2021).
11 S. 2992, 117th Cong. (2022) (reported with an amendment in the nature of a substitute) [hereinafter “Reported
Version”].
Congressional Research Service
1
link to page 24
The American Innovation and Choice Online Act (S. 2992)
released a different version of the AICOA on May 25, 2022.12 The May 25 draft indicates that
Senator Klobuchar intends to introduce it as an amendment in the nature of a substitute if the
Senate takes up the bil .
This report provides an overview of versions of the AICOA pending in the Senate. The report’s
discussion applies to both the reported version of S. 2992 and Senator Klobuchar’s draft
amendment, unless specifical y noted. While the report focuses on versions of the legislation
pending in the Senate, it contains several textboxes highlighting key differences between S. 2992
and H.R. 3816.
Covered Platforms
The AICOA would apply special rules to “covered platforms.” The May 25 draft defines that term
to mean “online platforms”13 that exceed certain thresholds for U.S.-based active users; exceed
certain thresholds for annual sales, market capitalization, or worldwide active users; and occupy
positions as “critical trading partners.”14
The bil would authorize the Department of Justice (DOJ) and Federal Trade Commission (FTC)
to jointly designate firms that meet these criteria as covered platforms.15 However, the legislation
does not limit its prohibitions to entities that have been formal y designated. Rather, the bil
appears to contemplate the possibility that regulators wil enforce its prohibitions against
non-designated entities that nevertheless qualify as covered platforms.16
Under the May 25 draft, the term “covered platform” would mean an online platform that:
12 T he May 25 version of the legislation is available at https://www.klobuchar.senate.gov/public/_cache/
files/b/9/b90b9806-cecf-4796-89fb-561e5322531c/B1F51354E81BEFF3EB96956A7A5E1D6A.sil22713.pdf
[hereinafter “May 25 Draft”].
13 T he May 25 version of the bill contains a definition of the term “online platform” that appears intended to exclude
financial-services and telecommunications companies from the legislation’s scope.
See May 25 Draft § 2(a)(9) (2022).
T he bill defines the term “online platform” to mean a website, online or mobile application, operating system, digital
assistant, or online service that “enables”:
A user to generate or share content that can be viewed by other users or to interact with other content on the
platform;
T he offering, advertising, sale, purchase, or shipping of products or services between and among consumers
or businesses not controlled by the platform operator; or
User searches or queries that access or display a volume of information.
Id. § 2(a)(9)(A). T he definition explicitly excludes services that provide the capability to transmit data to and receive
data from “all or substantially all internet endpoints” by wire or radio.
Id. § 2(a)(9)(B). In contrast, the reported version
of the bill would define the term “ online platform” to include websites, online or mobile applications, operating
systems, digital assistants, or online services that “
facilitate[] the offering, advertising, sale, purchase,
payment, or
shipping of products or services, including software applications, between and among consumers or businesses not
controlled by the platform operator.” S. 2992, 117th Cong. § 2(a)(9)(B) (Reported Version) (emphasis added). Unlike
the May 25 draft, the reported version of the bill would not specifically exclude telecommunications companies.
14 May 25 Draft § 2(a)(5)(B).
15 S. 2992, 117th Cong. § 3(d) (Reported Version); May 25 Draft § 3(d). For additional discussion of the designation
process and other issues related to the bill’s enforcement, see
“ Enforcement”
infra.
16
See, e.g., S. 2992, 117th Cong. § 2(a)(5)(B)(ii)(I) (Reported Version) (defining a “covered platform” to include
certain firms with large numbers of U.S.-based active users “ during the 12 months preceding a designation . . .
or the
12 months preceding the filing of a complaint for an alleged violation of this Act”) (emphasis added).
Congressional Research Service
2
The American Innovation and Choice Online Act (S. 2992)
Has at least 50 mil ion U.S.-based monthly active users or 100,000 U.S.-based
monthly active business users at any point during the 12 months preceding a
designation decision or the filing of a complaint for a violation of the bil ;
Is owned or control ed by an entity with:
Annual sales exceeding $550 bil ion at any point during the two years
preceding a designation decision or the filing of a complaint;
An average market capitalization exceeding $550 bil ion over any 180-day
period during the two years preceding a designation decision or the filing of a
complaint; or
At least one bil ion worldwide monthly active users during the 12 months
preceding a designation decision or the filing of a complaint; and
Is a “critical trading partner” for the sale or provision of any product or service
offered on or directly related to the platform.17
The AICOA defines the term “critical trading partner” to mean a person that has the ability to
“restrict or material y impede” a business user’s access to its users, customers, or a tool or service
needed to effectively serve its users or customers.18
Depending on the interpretation of the “critical trading partner” requirement, the legislation may
encompass a range of popular platforms, including:
Facebook (a social network), Instagram (a photo-sharing service), and WhatsApp
(a messaging application)—al of which are controlled by Meta Platforms;
Google Search (a search engine), YouTube (a video-sharing platform), Google
Ads (an online advertising platform), Android OS (a mobile operating system),
and the Google Play Store (a software application store)—al of which are
controlled by Alphabet;
Amazon Marketplace (an e-commerce marketplace) and Amazon Web Services
Marketplace (an online software store);
Apple’s App Store (a software application store) and iOS (a mobile operating
system);
Azure Marketplace (an online software store), LinkedIn (an employment-
oriented social network), Microsoft Store on Xbox (an online video-game store),
and Microsoft Windows (a group of operating systems)—al of which are
controlled by Microsoft; and
TikTok (a video-sharing service), which is controlled by the Chinese firm
ByteDance.19
The bil ’s supporters have argued that the criteria governing a firm’s status as a covered platform
are reasonable proxies for the type of “gatekeeping” power that often raises competition
17 May 25 Draft § 2(a)(5)(B). T he reported version of the bill would apply the same criteria as the May 25 version to
platforms controlled by publicly traded companies.
See S. 2992, 117th Cong. § 2(a)(5)(B) (Reported Version).
However, instead of the relevant sales and market -capitalization thresholds, the reported version would apply an
earnings threshold of $30 billion to platforms controlled by non -publicly traded companies.
Id. § 2(a)(5)(C) (Reported
Version).
18 S. 2992, 117th Cong. § 2(a)(6) (Reported Version); May 25 Draft § 2(a)(6).
19 Leah Nylen,
Tech Antitrust Bill Threatens to Break Apple, Google’s Grip on the Internet, BLOOMBERG (July 26,
2022), https://www.bloomberg.com/graphics/2022-tech-antitrust -bill/.
Congressional Research Service
3
The American Innovation and Choice Online Act (S. 2992)
concerns.20 Critics have contended that the thresholds are arbitrary and that a firm’s market
capitalization, annual sales, and user base have little relevance for its ability to harm
competition.21
The legislation’s “critical trading partner” requirement arguably addresses some of this concern
about arbitrariness.22 However, that requirement would need to be interpreted by the DOJ, the
FTC, and the courts. As noted, S. 2992 defines the term “critical trading partner” to mean an
entity with the ability to “restrict or material y impede” a business user’s access to its customers
or necessary inputs.23 The bil does not contain further clarification of this language, which is not
drawn from antitrust case law. Existing antitrust doctrine instead emphasizes
market power,
which is a requirement for most forms of antitrust liability.24 Plaintiffs typical y establish market
power by showing that a defendant occupies a large share of a properly defined antitrust market.25
In brief, defining an antitrust market involves an evaluation of the range of reasonable substitutes
for a given product or service.26
The relationship between S. 2992’s “critical trading partner” requirement and these standards
from existing doctrine is not entirely clear. The core concern of market-power analysis—the
availability of reasonable substitutes—seems relevant to whether a platform has the ability to
“restrict or material y impede” a business user’s access to customers or inputs. It is notable,
however, that the AICOA does not employ the familiar language of market power.
The choice to use new language that lacks an accepted meaning may reflect an intent to adopt less
demanding standards than those in the market-power case law. Some commentators have
expressed dissatisfaction with aspects of the relevant doctrine. For example, critics have argued
that the Supreme Court’s 2018 decision in
Ohio v. American Express27—which adopted special
market-definition rules for “two-sided” markets—may hamper antitrust enforcement against tech
20
See, e.g., Adam Conner & Erin Simpson,
Evaluating 2 Tech Antitrust Bills to Restore Competition Online, CTR. FOR
AM. PROGRESS (June 2, 2022), https://www.americanprogress.org/article/evaluating-2-tech-antitrust-bills-to-restore-
competition-online/.
21
See, e.g., Aurelien Portuese,
The Revised (But Uncorrected) Version of the Klobuchar Bill, INFO. T ECH. &
INNOVATION FDN. (June 21, 2022), https://itif.org/publications/2022/06/21/the-revised-but-uncorrected-version-of-the-
klobuchar-bill/.
22
See Scott Morton, et al.,
supra note 9, at 1 (arguing that the AICOA is “carefully targeted in that its prohibitions
apply only to platforms deemed ‘critical trading partners’”).
23 S. 2992, 117th Cong. § 2(a)(6) (Reported Version); May 25 Draft § 2(a)(6).
24
See, e.g., John B. Kirkwood,
Market Power and Antitrust Enforcement, 98 B.U. L. REV. 1169, 1173 (2018)
(explaining that market power “is central to antitrust because it distinguishes firms that can harm competition and
consumers from those that cannot”). A firm possesses market power when it has the ability to profitably charge prices
above competitive levels for a sustained period of time. MARKET POWER HANDBOOK: COMPETITION LAW AND
ECONOMIC FOUNDATIONS, AM. BAR ASS’N 1 (2d ed. 2012). Many economists contend that proper assessments of
market power involve quality considerations.
See id. at 1 n.3. T hus, market -power analysis may require inquiries into a
firm’s ability to degrade the quality of its offerings without facing a meaningful risk of consumer substitution.
See
generally John M. Newman,
Antitrust in Zero-Price Markets: Applications, 94 WASH. U. L. REV. 49 (2016); John M.
Newman,
Antitrust in Zero-Price Markets: Foundations, 164 U. PA. L. REV. 149 (2015).
25 MARKET POWER HANDBOOK,
supra note 24, at 18. Plaintiffs can also establish market power with direct evidence of
supra-competitive prices, but this can be a difficult task.
See id. at 20.
26
See id. at 61-63. More technically, market definition typically involves analysis of the cross-elasticity of demand
between different products—that is, the extent to which the quantity demanded of one product will change in response
to a change in the price of another product.
Id. at 64.
27 138 S. Ct. 2274 (2018).
Congressional Research Service
4
The American Innovation and Choice Online Act (S. 2992)
platforms.28 Other scholars have criticized the market-definition paradigm more general y.29
Senator Klobuchar—S. 2992’s sponsor—has introduced a different antitrust bil that appears to
reflect both of those concerns.30
This al suggests that the AICOA’s “critical trading partner” requirement is not intended to
incorporate current market-power doctrine wholesale. Nevertheless, the extent to which
preexisting antitrust principles would influence the interpretation of the “critical trading partner”
language remains an open question. One observer has characterized the bil as repudiating
any analysis of market power and criticized it on that basis.31 However, the legislation might instead
prompt courts to dispense with some of the more demanding elements of current doctrine without
completely abandoning certain general principles that inform it.
These types of interpretive issues recur throughout the bil and could present regulators and courts
with difficult questions if the legislation becomes law.
28 Lina M. Khan,
The Supreme Court Case That Could Give Tech Giants More Power, N.Y. T IMES (Mar. 2, 2018),
https://www.nytimes.com/2018/03/02/opinion/the-supreme-court-case-that-could-give-tech-giants-more-power.html.
29
See Louis Kaplow,
Why (Ever) Define Markets?, 124 HARV. L. REV. 437 (2010);
see also INVESTIGATION OF
COMP ETITION IN DIGITAL MARKETS,
supra note 1, at 399 (recommending that Congress enact legislation providing that
market definition is not required to prove an antitrust violation).
30
See S. 225, 117th Cong. §§ 9, 13 (2021).
31 Erik Hovenkamp,
Proposed Antitrust Reforms in Big Tech: What Do They Imply for Competition and Innovation?,
COMP ETITION POLICY INT’L ANTITRUST CHRONICLE 15, 22 (July 2022).
Congressional Research Service
5
The American Innovation and Choice Online Act (S. 2992)
Comparing Different Versions of the AICOA: S. 2992 and H.R. 3816
In June 2021, the House Judiciary Committee ordered the committee version of the AICOA—H.R. 3816—to be
reported with amendments to the ful House. While parts of H.R. 3816 overlap with S. 2992, the bil s also differ in
several respects.
Like S. 2992, H.R. 3816 would apply special rules to “covered platforms.” H.R. 3816, 117th Cong. § 2 (2021).
There is some similarity in the criteria used to determine a firm’s status as a covered platform: the bil s contain
the same thresholds for monthly active U.S. users and the same “critical trading partner” standard.
See S. 2992,
117th Cong. §§ 2(a)(5)(B)(i )(I), 2(a)(6) (2022) (Reported Version); May 25 Draft §§ 2(a)(5)(B)(i), 2(a)(6) (2022);
H.R. 3816, 117th Cong. §§ 2(g)(4)(B)(i), 2(g)(5).
There are differences with respect to the other criteria. Under H.R. 3816, the relevant sales and
market-capitalization thresholds would be $600 bil ion, as opposed to $550 bil ion u nder S. 2992.
See S. 2992,
117th Cong. § 2(a)(5)(B)(i )(II)(aa) (Reported Version); May 25 Draft § 2(a)(5)(B)(ii)(I); H.R. 3816, 117th
Cong. § 2(g)(4)(B)(i ).
Likewise, while S. 2992 could apply to entities that fal below the relevant sales and market-capitalization
thresholds if they have more than one bil ion worldwide monthly active users, H.R. 3816 would not.
See S. 2992,
117th Cong. § 2(a)(5)(B)(i )(II)(bb) (Reported Version); May 25 Draft § 2(a)(5)(B)(i )(II); H.R. 3816, 117th
Cong. § 2(g)(4)(B)(i ).
There are also two differences related to designation decisions. First, S. 2992 would al ow the DOJ and FTC to
jointly designate firms that meet the specified criteria as covered platforms, while H.R. 3816 appears to permit
either agency to do so independently. S. 2992, 117th Cong. § 3(d) (Reported Version); May 25 Draft § 3(d); H.R.
3816, 117th Cong. § 2(d). Second, designation decisions under S. 2992 would be valid for seven years, as opposed
to 10 years under the House committee-reported bil . S. 2992, 117th Cong. § 3(d)(1)(C) (Reported Version); May
25 Draft § 3(d)(1)(C); H.R. 3816, 117th Cong. § 2(d)(3).
Unlawful Conduct
The AICOA would prohibit covered platforms from engaging in 10 categories of conduct.32
Under the May 25 version of the legislation, firms that violate the bil ’s prohibitions would be
liable for up to 10% of their total U.S. revenue for the period in which the violation occurred.33
The bil makes “material harm” to competition an element of three of the offenses.34 For the
remaining seven categories, defendants would be al owed to rebut a
prima facie case by showing
that their conduct would
not result in “material harm” to competition.35 The legislation would also
offer several other affirmative defenses for al 10 offenses.36
This section provides an overview of the bil ’s prohibitions, their relationship to current antitrust
doctrine, and related interpretive issues. The bil ’s affirmative defenses are discussed later in the
report.37
32 S. 2992, 117th Cong. § 3(a)(1)-(10) (2022) (Reported Version); May 25 Draft § 3(a)(1)-(10) (2022).
33 May 25 Draft § 3(c)(6)(B). Under the reported version of S. 2992, violators would be liable for up to 15 percent of
their total U.S. revenue for the period in which the violation occurred. S. 2992, 117th Cong. § 3(c)(5)(B) (Reported
Version).
34 S. 2992, 117th Cong. § 3(a)(1)-(3) (Reported Version); May 25 Draft § 3(a)(1)-(3).
35 S. 2992, 117th Cong. § 3(b)(2)(A) (Reported Version); May 25 Draft § 3(b)(2).
36 S. 2992, 117th Cong. § 3(b)(1)-(2) (Reported Version); May 25 Draft § 3(b)(1).
37
See “Affirmative Defenses”
infra.
Congressional Research Service
6
The American Innovation and Choice Online Act (S. 2992)
Sections 3(a)(1)-(3): Self-Preferencing, Limitations on Business
Users, and Discrimination That Would “Materially Harm
Competition”
Sections 3(a)(1)-(3) of the AICOA would prohibit covered platforms from engaging in certain
forms of conduct in a manner that would “material y harm competition.”38
Section 3(a)(1) would prohibit operators of covered platforms from preferencing
their own products, services, or lines of business over those of other business
users of their platforms in a manner that would “material y harm competition.”39
Section 3(a)(2) would prohibit operators of covered platforms from limiting the
ability of business users to compete with the operators’ own offerings in a
manner that would “material y harm competition.”40
Section 3(a)(3) would prohibit operators of covered platforms from
discriminating in the application of their terms of service among similarly
situated business users in a manner that would “material y harm competition.”41
Much of the commentary surrounding the bil has focused on the types of conduct that may
trigger liability under these provisions. Some of the practices that commentators have flagged as
potential y prohibited include:
Google Search’s display of Google Maps content with search results, in addition
to its favorable placement of other Google verticals;42
Amazon advantaging its own products in search results on its Marketplace or
favoring itself over third-party merchants in managing its “Buy Box”;43
Apple’s preinstal ation of certain software applications (apps) on its mobile
devices;44 and
Microsoft’s prioritization of its own video games in Microsoft Store on Xbox.45
38 S. 2992, 117th Cong. § 3(a)(1)-(3) (Reported Version); May 25 Draft § 3(a)(1)-(3).
39 S. 2992, 117th Cong. § 3(a)(1) (Reported Version); May 25 Draft § 3(a)(1).
40 S. 2992, 117th Cong. § 3(a)(2) (Reported Version); May 25 Draft § 3(a)(2).
41 S. 2992, 117th Cong. § 3(a)(3) (Reported Version); May 25 Draft § 3(a)(3).
42 Neil Bradley,
Congressional Price Hike? How the American Innovation and Choice Online Act Raises Prices on
Consum ers, U.S. CHAMBER OF COMMERCE (June 21, 2022), https://www.uschamber.com/finance/antitrust/
congressional-price-hike-how-the-american-innovation-and-choice-online-act-raises-prices-on-consumers.
43 Hovenkamp,
supra note 7, at 1546-47; Nylen,
supra note 19. Amazon’s “Buy Box” is an icon that allows consumers
to either place items in their carts or buy them with a single click. T hird-party merchants compete with one another and
with Amazon to “win” the Buy Box for a given product at any point in time. Winners are chosen based on factors that
include price and reliability in fulfillment.
See Brian Connolly,
How to Win the Am azon Buy Box in 2021 , JUNGLE
SCOUT (Jan. 4, 2021), https://www.junglescout.com/blog/how-to-win-the-buy-box/.
44 Randy Picker,
How Would the Big Tech Self-Preferencing Bill Affect Users?, PROMARKET (June 16, 2022),
https://www.promarket.org/2022/06/16/how-would-the-big-tech-self-preferencing-bill-affect -users/;
but see Hal Singer,
Rep. Cicilline’s Nondiscrim ination Bill Would Offer a Lifeline to Independent App Developers, PROMARKET (July 2,
2021), https://www.promarket.org/2021/07/02/antitrust -self-preferencing-preinstallation-app-developers-apple/
(arguing that similar language in the House version of the AICOA would not prohibit app preinstallation).
45 Nylen,
supra note 19.
Congressional Research Service
7
The American Innovation and Choice Online Act (S. 2992)
The House subcommittee’s October 2020 report on digital competition identified several of these
practices as examples of al egedly anticompetitive conduct by tech platforms.46 Whether S. 2992
would in fact proscribe this conduct would depend on the application of key terms like
“preference,” “limit,” and “materially harm competition.”
In addition to these behaviors—which principal y advantage a platform operator over its rivals—
the legislation may bar certain forms of “secondary line” discrimination.47 In particular, some
commentators have argued that Section 3(a)(3)’s prohibition of discriminatory terms-of-service
enforcement may limit firms’ ability to remove unwanted content from their platforms.48
The following subsections review these issues.
Self-Preferencing
The bil ’s prohibition of self-preferencing may raise particularly complex questions. Many forms
of conduct by a vertical y integrated firm arguably “preference” the firm’s own offerings over
those of its rivals. Interpreted literal y, that language could encompass any actions that treat a
platform’s own products more favorably than those of competitors.49 If the term “preference” is
given this type of expansive meaning, Section 3(a)(1)’s “material y harm competition”
requirement would likely do much of the work defining the provision’s boundaries.
At the same time, the concept of self-preferencing may also have internal limits that could
complicate the bil ’s enforcement. Several commentators have argued that proving “preferential”
conduct would require regulators to establish some baseline level of neutral treatment from which
a defendant has deviated.50 In this view, self-preferencing would not encompass a platform
operator’s favorable treatment of its own offerings when they
deserve such treatment (
i.e., when a
platform operator’s offerings are better than the alternatives).51
Section 3(a)(1) may thus require regulators and courts to grapple with chal enging conceptual
questions.52 While cases in which a firm overrides its ordinary algorithmic protocols to
disadvantage rivals may be fairly straightforward instances of self-preferencing, other fact
patterns could raise thornier interpretive and evidentiary issues.53
46 INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 14 (concluding that Google has “used its
search monopoly . . . to boost Google’s own inferior vertical offerings, while imposing search penalties to demote
third-party vertical providers”);
id. at 282 (concluding that Amazon has given itself “ favorable treatment relative to
competing sellers” through its control of the Buy Box);
id. at 352-54 (discussing Apple’s preinstallation of its own apps
on iPhones).
47 “Secondary line” competitive injury occurs when a company’s discriminatory conduct disadvantages some of its
customers relative to others in a manner that harms competition.
See HOVENKAMP,
supra note 7, at 632.
48
See note 82
infra.
49
See Hovenkamp,
supra note 31.
50 For versions of this argument, see T homas A. Lambert,
Addressing Big Tech’s Market Power: A Comparative
Institutional Approach, 75 SMU L. REV. 73, 98-99 (2022); Daniel A. Crane,
Search Neutrality as an Antitrust
Principle, 19 GEO. MASON L. REV. 1199 (2012); Geoffrey A. Manne & Joshua D. Wright,
If Search Neutrality is the
Answer, What’s the Question?, 2012 COLUM. BUS. L. REV. 151 (2012).
51 Lambert,
supra note 50, at 98-99. A product could conceivably be “better” than the alternatives in th e sense that it is
generally superior, or in the narrower sense that it works better with a platform operator’s other offerings. For example,
Apple might argue that its App Store is “better” than rival stores in the sense that the App Store is categorically
superior, or in the sense that it works better on the iOS operating system.
52
See generally Lawrence Solum,
Legal Theory Lexicon: Baselines, LEGAL THEORY BLOG (Mar. 2, 2009),
https://lsolum.typepad.com/legaltheory/2009/03/legal-theory-lexicon-baselines.html.
53
See Crane,
supra note 50, at 1207-08.
Congressional Research Service
8
The American Innovation and Choice Online Act (S. 2992)
Material Harm to Competition
The “material y harm competition” requirement in Sections 3(a)(1)-(3) has also generated
discussion.
The AICOA’s supporters have identified that requirement as a key limiting principle in response
to al egations of overbreadth.54
The bil ’s opponents have argued that there are two senses in which the meaning of the
“material y harm competition” standard is unclear. First, the term “material y” does not have an
established meaning in antitrust doctrine. While Section 7 of the Clayton Act prohibits mergers
and acquisitions that may “substantial y” lessen competition,55 neither the antitrust statutes nor
the case law employ S. 2992’s language of materiality. The legislation’s critics have argued that
this novelty is likely to create significant uncertainties about the bil ’s application.56 S. 2992’s
proponents have answered that the new terminology is intended to strengthen existing law, which
they argue has proven inadequate to grapple with anticompetitive conduct by tech platforms.57
Second, some commentators have argued that the “material y harm competition” standard may be
interpreted in ways that depart from current understandings of competitive harm.58
One source of this concern involves differences between S. 2992 and H.R. 3816. The “material y
harm competition” standard in S. 2992 differs from language in H.R. 3816 that offers defendants
an affirmative defense for conduct that does not harm “the competitive process by restricting or
impeding legitimate activity by business users.”59 Critics have contended that this contrast
suggests “a shift away from protecting competitive processes towards protecting individual
competitors.”60
Another observer has argued that the AICOA’s omission of certain traditional antitrust concepts
like market power may prompt courts to conclude that the bil seeks to move beyond preexisting
notions of harm to competition.61
54
See, e.g., Aaron Schur,
The Critiques Against the American Innovation and Choice Online Act Miss the Mark,
PROMARKET (July 18, 2022), https://www.promarket.org/2022/07/18/the-critiques-against-the-american-innovation-
and-choice-online-act-miss-the-mark/; Scott Morton, et al.,
supra note 9, at 6; Bill Baer,
Why Am azon is Wrong About
the Am erican Innovation and Choice Online Act, BROOKINGS INST. (June 14, 2022), https://www.brookings.edu/
blog/techtank/2022/06/14/why-amazon-is-wrong-about -the-american-innovation-and-online-choice-act/; Conner &
Simpson,
supra note 20.
55 15 U.S.C. § 18. Senator Klobuchar has introduced other legislation that would amend Section 7 of the Clayton Act to
prohibit mergers and acquisitions that “create an appreciable risk of materially lessening competition.” S. 225, 117th
Cong. § 4(b) (2021). T hat legislation defines the term “materially” to mean “more than a de minimis amount.”
Id. S.
2992 leaves the term “ materially” undefined.
56
See, e.g., Comments of the Am. Bar Ass’n Antitrust L. Section Regarding the Am. Innovation and Choice Online Act
(S. 2992) Before the 117th Congress at 9 (Apr. 27, 2022), https://www.americanbar.org/content/dam/aba/
administrative/antitrust_law/comments/at -comments/2022/comments-aico-act.pdf [hereinafter “ ABA Comments”].
57 Scott Morton, et al.,
supra note 9, at 2 (characterizing the novelty of the “materially harm competition” standard as a
“feature, not a bug” of the AICOA); Conner & Simpson,
supra note 20.
58 A. Douglas Melamed,
Why I Think Congress Should Not Enact the American Innovation and Choice Online Act,
COMP ETITION POLICY INT’L (June 19, 2022), https://www.competitionpolicyinternational.com/why -i-think-congress-
should-not-enact -the-american-innovation-and-choice-online-act/; Portuese,
supra note 21; ABA COMMENTS,
supra note 56, at 5-6, 9, 20.
59 H.R. 3816, 117th Cong. § 2(c)(1)(A) (2021).
60 ABA COMMENTS,
supra note 56, at 5.
61 Melamed,
supra note 58.
Congressional Research Service
9
The American Innovation and Choice Online Act (S. 2992)
Some of the AICOA’s supporters disagree. Representative David Cicil ine—H.R. 3816’s
sponsor—has contended that S. 2992’s reference to harm to “competition” incorporates that
concept’s established meaning in the case law, which encompasses harm to the “competitive
process” but not harm to individual competitors standing alone.62 In this view, the difference
between H.R. 3816’s reference to harm to the “competitive process” and S. 2992’s notion of harm
to “competition” is stylistic rather than substantive.
While harm to “competition” is a central concept in antitrust jurisprudence, its precise meaning is
contested in multiple ways. There is a long-standing debate over whether various goals—like
protecting “consumer welfare,” the “competitive process,” or “trading partner welfare”—
accurately reflect the meaning of “competition” that has developed in the antitrust case law.63
Some if not al of those goals themselves have several possible meanings.64 Final y, in addition to
these doctrinal debates, there are differences of opinion over the types of harm to “competition”
that antitrust
should preclude.65
It is unclear whether these debates would play a role in the interpretation of S. 2992’s “material y
harm competition” standard. If the bil became law, courts may conclude that the standard ratifies
prevailing notions of competitive harm, whatever the courts determine those to be. To similar
effect, judges steeped in the reigning approach to antitrust may be hesitant to embrace innovative
theories of harm without more explicit legislative direction.66
There are possible counterarguments, however. Recent efforts to reorient antitrust’s focus might
support the conclusion that the meaning of harm to “competition” in S. 2992 is not identical to
preexisting understandings of that concept in the case law. For example, the October 2020 House
subcommittee report on digital competition that laid the groundwork for the AICOA concluded
that the Supreme Court “has limited the analysis of competitive harm to focus primarily on price
and output rather than the competitive process.”67 In place of the current approach, the report
endorsed a broader standard that protects “not just consumers, but also workers, entrepreneurs,
independent businesses, open markets, a fair economy, and democratic ideals.”68 This diagnosis
and prescription could suggest an intent to reform and not merely ratify existing concepts of
competitive harm.
Several Members of Congress have noted these theoretical debates and expressed uncertainty
about the role that “consumer welfare” would play in the AICOA’s application.69 These concerns
62 Letter from Rep. David N. Cicilline to Hon. Brian Schatz, et al., at 2 (June 15, 2022), https://cicilline.house.gov/sites/
evo-subsites/cicilline.house.gov/files/evo-media-document/2022-0615-cicilline-letter-to-senators-re-content-
moderation.pdf [hereinafter “ Cicilline Letter”] (“ The antitrust concept of ‘harm to competition’ exists under current
law and means ‘harm, not just to a single competitor, but to the competitive process,
i.e., to competition itself.’”)
(quoting NYNEX Corp. v. Discon, Inc., 525 U.S. 128, 135 (1998)).
63
See, e.g., Herbert Hovenkamp,
The Slogans and Goals of Antitrust Law, U. Penn. Inst. for L. & Econ. Research
Paper No. 22-33 (Oct. 2, 2022), https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4121866# .
64
See id. at 49-89.
65
See, e.g., Barak Orbach,
How Antitrust Lost Its Goal, 81 FORDHAM L. REV. 2253 (2013); Maurice E. Stucke,
Reconsidering Antitrust’s Goals, 53 B.C. L. REV. 551 (2012). T here is also a dispute over whether antitrust needs a
more specific normative benchmark than general n otions like “ competition.”
See, e.g., Eleanor M. Fox,
Against Goals,
81 FORDHAM L. REV. 2157 (2013).
66
See Scott Morton, et al.,
supra note 9, at 3 (“[A] judiciary weighted with years of experience and jurisprudence will
have significant inertia, and it is therefore unduly optimistic to imagine outcomes under the [AICOA] would veer
drastically away from past understandings of core concepts like harm to competition.”).
67 INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 391.
68
Id. at 392.
69
See, e.g., T ranscript of Markup of S. 2992 at 53 (Jan. 20, 2022) (on file with author) [hereinafter “S. 2992 Markup
Congressional Research Service
10
The American Innovation and Choice Online Act (S. 2992)
led the House Judiciary Committee to adopt an amendment to H.R. 3816 that provides an
affirmative defense for conduct that “increases consumer welfare.”70 S. 2992, in contrast, does not
contain a paral el affirmative defense. The role that consumer benefits would play in the
application of the “material y harm competition” standard thus remains unclear. If consumer
benefits and other efficiencies would not register in the application of that standard, then the
meaning of the term “competition” in S. 2992 would diverge from the meaning that many courts
ascribe to that concept under current law.71
Efforts to equate harm to “competition” with harm to the “competitive process” do not
necessarily resolve such ambiguities.72 The “competitive process” has been invoked for a variety
of purposes, some of which are in tension with one another. For example, that phrase has been
used:
To distinguish harm to “competition” (which qualifies as a cognizable harm
under current antitrust law) from harm to individual competitors (which does
not);73
To refer to a general standard under which antitrust is or should be concerned
with the protection of competition as a process rather than with the achievement
of specific economic outcomes;74 and
T ranscript”] (Sen. T hom T illis) (“It’s not clear how existing competitor or competition jurisprudence would support or
be changed by [S. 2992]. T he purpose of competition law is to eliminate harm to consumers not to pick winners and
losers. I’m also aware of the spirited debate [over] whether decades of antitrust law based on [the] consumer-welfare
standard should be put in the burn pit. I’m open to having [a] separate discussion about potential changes to that
standard and I hope that we will. But as it stands in relation to this bill, what standard will enforcers look to [?] What
about amendments [that] would insert [the] consumer welfare standard back into the definition of material harm to
competition?”); T ranscript of Markup of H.R. 3843, the Merger Filing Fee Modernization Act, et al., at 16,866 -16,869
(June 23, 2021), https://www.congress.gov/117/meeting/house/112818/documents/HMKP -117-JU00-Transcript-
20210623.pdf [hereinafter “ H.R. 3816 Markup T ranscript”] (Rep. Cliff Bentz) (“ [W]e should also take into [account]
consumers, since that is the proper focus and long-time focus, over the past 40 years at least, of our antitrust law.”);
id.
at 16,887-16,888 (Rep. Zoe Lofgren) (“ [I]n this bill there’s no consideration whatsoever for consumer benefit.”).
70 H.R. 3816, 117th Cong. § 2(c) (2021); H.R. 3816 Markup T ranscript at 17,309-17,310.
71
See, e.g., United States v. Microsoft, 253 F.3d 34, 59 (D.C. Cir. 2001) (explaining that if a plaintiff makes a
prima
facie case of anticompetitive harm under Section 2 of the Sherman Act, the defendant may proffer a procompetitive
justification for its conduct —that is, “ a nonpretextual claim that its conduct is indeed a form of competition on the
merits because it involves, for example, greater efficiency or enhanced consumer appeal”);
see also NCAA v.
Alston, 141 S. Ct. 2141, 2151 (2021) (explaining that in Rule-of-Reason cases under Section 1 of the Sherman Act, the
goal is “[a]lways” to “distinguish between restraints [of trade] with anticompetitive effect that are harmful to the
consumer and restraints stimulating competition that are in the consumer’s best interest”) (internal quotation marks and
citation omitted).
72
See Hovenkamp,
supra note 63, at 51 (“The ‘competitive process’ can mean pretty much what anyone thinks it
means. As a result it embraces mutually inconsistent antitrust ideologies.”); John M. Newman,
Procom petitive
Justifications in Antitrust Law, 94 IND. L.J. 501, 514 (2019) (“[T]he actual content of the competitive-process approach
remains mercurial, a cipher . . . . More than a half-century has passed since the [Supreme] Court first clearly invoked
the competitive process approach to condemn a restraint of trade, yet terms like ‘competition’ and ‘competitive
process’ are still ‘wonderfully ill-defined.’”) (citation omitted); Stucke,
supra note 65, at 569 (arguing that the
“competitive process” standard fails to provide meaningful guidance because “it simply shifts the debate [over
antitrust’s goals] to a larger, unresolved issue, namely defining an ‘effective competitive process,’” and “[n]o
consensus exists in the United States or elsewhere on an effective competition process or a unifying theory of
competition.”).
73
E.g., Broadcom Corp. v. Qualcomm, Inc., 501 F.3d 297, 308 (3d Cir. 2007).
74
E.g., T im Wu,
The “Protection of the Competitive Process” Standard, Colum. Pub. L. Research Paper No. 14-612
(2018), https://scholarship.law.columbia.edu/cgi/viewcontent.cgi?article=3293&context=faculty_scholarship ; Lina M.
Khan,
Am azon’s Antitrust Paradox, 126 YALE L.J. 710, 745 (2017); Gregory J. Werden,
Antitrust’s Rule of Reason:
Congressional Research Service
11
The American Innovation and Choice Online Act (S. 2992)
In ways that are difficult to distinguish from canonical statements of the
“consumer welfare” standard.75
Accordingly, courts would likely face a range of plausible meanings in interpreting and applying
S. 2992’s “material y harm competition” standard, even if that standard is equivalent to H.R.
3816’s reference to harm to the “competitive process.”76
In sum, then, the “material y harm competition” standard would have to be fleshed out in
practice. The AICOA would direct the DOJ and FTC to issue guidelines outlining their
interpretation of that language.77 Those guidelines could give regulated entities some clarity as to
the standard’s scope,78 but would not be legal y binding.79 As a result, the judiciary would likely
unpack the legislation’s meaning over time, presumably operating under the assumption that the
“material y harm competition” standard is intended to be more restrictive than certain aspects of
general antitrust doctrine.80
Only Com petition Matters, 79 ANTITRUST L.J. 713, 732 (2014); Eleanor M. Fox,
Modernization of Antitrust: A New
Equilibrium , 66 CORNELL L. REV. 1140, 1152-54 (1981).
75
E.g., T own of Concord v. Boston Edison Co., 915 F.2d 17, 21 -22 (1st Cir. 1990) (Breyer, J.) (stating that a practice
harms the “competitive process” if it “obstructs the achievement of competition’s basic goals—lower prices, better
products, and more efficient production methods”);
see also Lina M. Khan,
The Ideological Roots of America’s Market
Power Problem , YALE L.J. F. 960, 971 n.48 (2018) (“ Even when courts claim to be analyzing the competitive process,
they tend to use consumer welfare as a proxy to assess it.”) (citing United States v. Microsoft Corp., 253 F.3d 34, 58
(D.C. Cir. 2001)).
Like the “competitive process,” “consumer welfare” has several possible meanin gs.
See, e.g., Stucke,
supra note 65, at
571-73; Barak Orbach,
The Antitrust Consum er Welfare Paradox, 7 J. COMPETITION L. ECON. 133 (2011). One scholar
has explained that “under the consumer welfare . . . principle as most people understand it today,” antitrust “encourages
markets to produce output as high as is consistent with sustainable competition, and prices that are accordin gly as low.”
Herbert J. Hovenkamp,
Is Antitrust’s Consum er Welfare Principle Im periled? , 45 J. CORP . L. 101, 102 (2019). Some
commentators have argued that —properly understood—the “ consumer welfare” standard incorporates non -price
benefits to consumers, such as product quality and innovation.
See, e.g., Christine S. Wilson,
Welfare Standards
Underlying Antitrust Enforcem ent: What You Measure is What You Get, Luncheon Keynote Address at Geo. Mason L.
Rev. 22nd Annual Antitrust Symposium at 5-7 (Feb. 15, 2019), https://www.ftc.gov/system/files/documents/
public_statements/1455663/welfare_standard_speech_-_cmr-wilson.pdf.
76 Disputes over whether the “competitive process” standard is already the law may further complicate the interpretive
task. Some commentators have argued that the “competitive process” approach represents an accurate statement of
current antitrust doctrine.
E.g., Werden,
supra note 74, at 732. Others have been explicit that they regard the
“competitive process” standard as a desirable
replacement for prevailing understandings of competitive harm in the
case law.
E.g., INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 391-92; Wu,
supra note 74;
Khan,
supra note 74, at 745;
see also Einer Elhauge,
Should the Com petitive Process Test Replace the Consum er
Welfare Standard?, PROMARKET (May 24, 2022), https://www.promarket.org/2022/05/24/should-the-competitive-
process-test-replace-the-consumer-welfare-standard/ (criticizing efforts to replace the “ consumer welfare” standard
with the “competitive process” standard).
77 S. 2992, 117th Cong. § 4 (2022) (Reported Version); May 25 Draft § 4 (2022).
78 Scott Morton, et al.,
supra note 9, at 2 (“Any uncertainty about the meaning of words like ‘competition’ will be
resolved in [the agencies’] . . . guidelines and over time with the development of caselaw.”).
79 S. 2992, 117th Cong. § 4(d) (Reported Version); May 25 Draft § 4(d).
80 In particular, Sections 3(a)(1)-(3) seem intended to repudiate the permissive rules that currently govern unilateral
refusals to deal.
See Verizon Commc’ns Inc. v. Law Offs. of Curtis V. T rinko, LLP, 540 U.S. 398 (2004). T hose rules
would likely apply to much of the conduct targeted by Sections 3(a)(1) -(3), if such conduct was challenged under
existing law. For an overview of this case law, see Hovenkamp,
supra note 7, at 1495-1502. By subjecting the specified
forms of conduct to a general standard of competitive harm, the AICOA appears to dispense with such conduct -specific
tests.
Congressional Research Service
12
The American Innovation and Choice Online Act (S. 2992)
Content Moderation
Commentators have also debated Section 3(a)(1)(3)’s possible effects on content moderation by
covered platforms. As discussed, that provision would prohibit operators of covered platforms
from discriminating in the enforcement of their terms of service among similarly situated
business users in a manner that would “material y harm competition.”81
Some observers have concluded that this language may restrict firms’ ability to remove unwanted
content from their platforms.82 To those concerned about al eged discrimination against
conservative viewpoints, this would be a virtue.83 To some of those worried about discouraging
tech platforms from policing hate speech and disinformation, it is a cause of unease.84
Whether Section 3(a)(1)(3) would in fact prohibit certain forms of content moderation is unclear.
As discussed, the “material y harm competition” standard is subject to differing interpretations,
making it difficult to confidently predict whether it would reach discriminatory content
moderation.
Several commentators have argued that it is unlikely that courts would find that
content-moderation decisions have the requisite anticompetitive effects.85 During H.R. 3816’s
markup, Representative Cicil ine also repeatedly stated that the bil is not intended to affect
content moderation.86
Others have argued that it is plausible that courts would read the AICOA to apply to
discriminatory content moderation and have raised concerns that the bil may chil content
moderation regardless.87 Additional y, during H.R. 3816’s markup, multiple Members stated or
81 S. 2992, 117th Cong. § 3(a)(1)(3) (Reported Version); May 25 Draft § 3(a)(1)(3).
82
See, e.g., Letter from Sen. Brian Schatz, et al., to Sen. Amy Klobuchar (June 14, 2022), https://www.freepress.net/
sites/default/files/2022-06/senate-letter-asking-klobuchar-for-content-moderation-fix-on-s2992.pdf; Jane Bambauer &
Anupam Chander,
Bills Meant to Check Big Tech’s Power Could Lead to More Disinform ation, WASH. POST (June 6,
2022), https://www.washingtonpost.com/outlook/2022/06/06/antitrust -bills-big-tech-hate-speech-disinformation/;
Maria Curi,
Tech Antitrust Bill Stokes Content-Moderation Worries, BLOOMBERG LAW (May 12, 2022),
https://news.bloomberglaw.com/privacy-and-data-security/tech-antitrust-bill-stokes-lawmaker-content-moderation-
worries.
83
See, e.g., Michael R. Davis,
Conservatives Must Use Antitrust to Rein In Big Tech Monopolies, FOX NEWS (July 12,
2022), https://www.foxnews.com/opinion/conservatives-must-use-antitrust -rein-big-tech-monopolies.
84
See, e.g., Bambauer & Chander,
supra note 82; Mark MacCarthy,
Two Ways to Improve Senator Klobuchar’s
Needed Antitrust Legislation, BROOKINGS INST. (Feb. 8, 2022), https://www.brookings.edu/blog/techtank/
2022/02/08/two-ways-to-improve-senator-klobuchars-needed-antitrust -legislation/.
85
See, e.g., Schur,
supra note 54; Gilad Edelman,
The Weak Argument Jeopardizing Tech Antitrust Legislation , WIRED
(June 16, 2022), https://www.wired.com/story/american-innovation-choice-online-act-democrats-argument/; Cicilline
Letter,
supra note 62, at 2; Conner & Simpson,
supra note 20.
86 H.R. 3816 Markup T ranscript at 15,687-15,689, 16,077-16,079.
During that markup, the House Judiciary Committee rejected a proposed amendment that would have explicitly
insulated platforms from liability for certain types of content moderation.
See id. at 16,813-16,814; Amendment to the
Amendment in the Nature of a Substitute to H.R. 3816 Offered by Ms. Lofgren of California (June 23, 2021),
https://docs.house.gov/meetings/JU/JU00/20210623/112818/BILLS-117-HR3816-L000397-Amdt-4.pdf.
Some Members who voted against the amendment explained that they believed it was unnecessary because the bill
does not involve content moderation.
See H.R. 3816 Markup T ranscript at 16,071-16,072 (Rep. David Cicilline).
Others voted against the amendment because of concerns that platforms would use the amendment’s protections to
discriminate against conservative viewpoints.
See H.R. 3816 Markup T ranscript at 16,592-16,601 (Rep. Matt Gaetz).
87
See, e.g., Letter from Samir Jain, et al., to Sen. Amy Klobuchar and Sen. Charles Grassley (Mar. 9, 2022),
https://cdt.org/wp-content/uploads/2022/03/S2292-CDT -letter-3-9-22-FINAL.pdf.
During S. 2992’s markup, Senator Alex Padilla raised concerns that the bill may be const rued to prohibit certain types
Congressional Research Service
13
The American Innovation and Choice Online Act (S. 2992)
suggested that they interpreted the bil to prohibit discriminatory content moderation involving
business users.88
The legislation’s interaction with Section 230 of the Communications Decency Act and the First
Amendment—both of which insulate firms from liability for certain forms of content
moderation—may also raise complicated legal questions.89 For example, Representative Cicil ine
has argued that Section 5(2) of S. 2992—which provides that the legislation shal not be
construed to limit the application of “any law”90—would preserve Section 230’s protections for
content-moderation decisions.91 Other commentators have responded that Section 5(2) does not
resolve concerns about the AICOA’s effect on content moderation because the bil may prompt
courts to adopt narrow interpretations of Section 230.92 A full discussion of this debate is beyond
the scope of this report.
Sections 3(a)(4) and 3(a)(7): Interoperability and Access to Data
Section 3(a)(4) of the bil would impose interoperability requirements on covered platforms. The
provision would make it unlawful for operators of such platforms to “material y restrict, impede,
or unreasonably delay” a business user’s ability to access or interoperate with features that are
available to the operators’ own competing products or services.93
Section 3(a)(7) would address the related issue of access to data generated by business users on
covered platforms. The provision would make it unlawful for operators of covered platforms to
of content moderation.
See S. 2992 Markup T ranscript at 61-62 (“ [T]his bill may hamper the efforts of platforms to
address the spread of hate speech and misinformation and disinformation efforts online that have caused so many
recent problems for our democracy.”). Senator T ed Cruz also expressed support for the legislation based on his belief
that it would limit discriminatory content moderation.
Id. at 65 (“ I believe the bill would make some positive
improvement as it concerns the problem of censorship . . . . [T he bill] would provide protections to content providers
[and] to businesses that are discriminated against because of the content of what they pr oduce.”).
88 During H.R. 3816’s markup, the House Judiciary Committee considered a proposed amendment that would have
granted a private right of action to persons whose content is subject to “politically biased content moderation” by a
covered platform operator. H.R. 3816 Markup T ranscript at 15,598-15,982; Amendment to the Amendment in the
Nature of a Substitute to H.R. 3816 Offered by [Mr. Jordan of Ohio] (June 23, 2021),
https://docs.house.gov/meetings/JU/JU00/20210623/112818/BILLS-117-HR3816-J000289-Amdt -3.pdf.
Representative Zoe Lofgren contended that this amendment was unnecessary because the bill already prohibited
covered platform operators from discriminating among similarly situated business users. H.R. 3816 Markup T ranscript
at 15,651-15,655, 15,657-15,663. Representative Jim Jordan—the amendment’s sponsor—responded that the
amendment was intended to extend the legislation’s prohibition of discriminatory content moderation to encompass
discrimination involving individuals as well as business users.
Id. at 15,664-15,665, 15,674, 15,790-15,794. T he
proposed amendment was not adopted.
Id. at 15,981-15,982.
89 For an overview of Section 230 of the Communications Decency Act, see CRS Report R46751,
Section 230: An
Overview, by Valerie C. Brannon and Eric N. Holmes. For a discussion of the First Amendment issues involved in
regulating content moderation by social-media companies, see CRS Report R45650,
Free Speech and the Regulation of
Social Media Content, by Valerie C. Brannon; CRS Legal Sidebar LSB10618,
Trial Court Rules State Social Media
Law Likely Unconstitutional, by Valerie C. Brannon.
90 S. 2992, 117th Cong. § 5(2) (2022) (Reported Version); May 25 Draft § 5(2) (2022).
91 Cicilline Letter,
supra note 62, at 2.
92 Letter from T echFreedom, et al., to Hon. Brian Schatz, et al. (June 27, 2022), https://techfreedom.org/wp-
content/uploads/2022/06/2992-Response-to-Cicilline-letter.pdf.
93 S. 2992, 117th Cong. § 3(a)(4) (Reported Version); May 25 Draft § 3(a)(4). Unlike the repo rted version of the bill,
the May 25 draft contains an exemption for cases in which such access would lead to a “significant cybersecurity risk.”
May 25 Draft § 3(a)(4).
Congressional Research Service
14
link to page 23
The American Innovation and Choice Online Act (S. 2992)
“material y restrict or impede” a business user from accessing data it generates on such platforms
or data that platform users generate by interacting with a business user’s products or services.94
Like the remaining prohibitions discussed below, these provisions would not require regulators to
establish that a platform’s conduct harmed competition. Rather,
absence of competitive harm
would be an affirmative defense. Defendants could rebut a
prima facie case under
Sections 3(a)(4) and 3(a)(7) by establishing that their conduct had not and would not result in
“material harm to competition.”95 The legislation would also offer several other affirmative
defenses involving user privacy, data security, and preservation of platform functionality that are
discussed below.96
Interoperability requirements like Section 3(a)(4) are a response to the network effects that give
large tech platforms powerful incumbency advantages.97 In markets characterized by strong
network effects, a platform’s services become more valuable as more people use the platform.98
Network effects can thus operate as entry barriers that make it difficult for newcomers to compete
with an established firm. Such effects create a dilemma: upstart tech platforms may not be as
attractive to consumers as an incumbent because upstarts lack similarly large user bases, which
they cannot attain without being attractive to consumers.99 This phenomenon makes many digital
markets susceptible to “tipping”—once one platform reaches a certain scale, its network
advantages reinforce themselves and cement its dominance.100
In theory, interoperability mandates may be able to mitigate these effects.101 For example, such
requirements could prevent a covered platform from cutting off rival app developers from the
tools needed to access their platforms—a strategy that a powerful firm could use to preserve its
dominant position or leverage that position into adjacent markets.102
Broader interoperability duties could also facilitate competition by al owing consumers who do
not directly use a dominant network to indirectly benefit from its scale. For example, if
interoperability requires Facebook to al ow users of other social networks to communicate with
its users, then upstart social networks may find it easier to attract customers.103 Whether
94 S. 2992, 117th Cong. § 3(a)(7) (Reported Version); May 25 Draft § 3(a)(7).
95 S. 2992, 117th Cong. § 3(b)(2)(A) (Reported Version); May 25 Draft § 3(b)(2).
96
See “Affirmative Defenses” infra.
97
See Michael Kades & Fiona Scott Morton,
Interoperability as a Competition Remedy for Digital Networks, WASH.
CTR. FOR EQUITABLE GROWTH 1-2 (Sept. 23, 2020).
98 DENNIS W. CARLTON & JEFFREY M. PERLOFF, MODERN INDUSTRIAL ORGANIZATION 392-93 (4th ed. 2005).
99
See REPORT, COMM. FOR THE STUDY OF DIGITAL PLATFORMS, MARKET STRUCTURE AND ANTITRUST SUBCOMM.,
GEORGE J. STIGLER CTR. FOR THE STUDY OF THE ECONOMY AND THE STATE, UNIV. OF CHI. BOOTH SCH. OF BUS. 17
(July 1, 2019), https://research.chicagobooth.edu/-/media/research/stigler/pdfs/market -structure-report.pdf.
100 Kades & Scott Morton,
supra note 97, at 7-9; UNLOCKING DIGITAL COMPETITION, REPORT OF THE DIGITAL
COMPETITION EXPERT PANEL, HER MAJESTY’S TREASURY 4 (Mar. 2019), https://assets.publishing.service.gov.uk/
government/uploads/system/uploads/attachment_data/file/785547/unlocking_digital_competition_furman_review_web.
pdf. In such cases, competitive pressures must come from outside the relevant market —for example, from firms that
create new markets via disruptive technologies. In the jargon, competitive dynamics in tipped markets shift from
“competition
in the market” to “competition
for the market,” which may not be forthcoming for extended periods of
time. UNLOCKING DIGITAL COMPETITION,
supra note 100, at 4, 38-41.
101 Kades & Scott Morton,
supra note 97, at 12.
102 T he FT C has alleged that Facebook used this strategy when it withheld critical tools from certain rival app
developers, including a rival social network called Path.
See FT C v. Facebook, Inc., 560 F. Supp. 3d 1, 6 (D.D.C.
2021). A House subcommittee also found that Apple has denied certain application programming interfaces and device
functionalities to competing developers. INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 354.
103 Kades & Scott Morton,
supra note 97, at 11-12.
Congressional Research Service
15
link to page 23
The American Innovation and Choice Online Act (S. 2992)
Section 3(a)(4) would entail such duties would depend on the details surrounding its
implementation.
Section 3(a)(7)’s data-access requirement would address related issues. By al owing business
users to access data involving their activities and customers on covered platforms, the provision
could mitigate the switching costs that can diminish consumers’ wil ingness to abandon an
incumbent platform in favor of a smal er rival.104
Interoperability requirements also have their skeptics, however. Depending upon their details,
such requirements may prove costly to implement.105 They may also dampen investment
incentives by forcing firms to share the fruits of their innovation with rivals.106 Additional y,
specific types of interoperability may create privacy concerns and data-security risks.107 While the
AICOA’s affirmative defenses attempt to address these latter concerns, defendants would bear the
burden of establishing their applicability.108
Section 3(a)(5): Tying
Section 3(a)(5) of the legislation would prohibit certain tying arrangements, whereby firms offer
one product on the condition that customers purchase a separate product as wel . The provision
would make it unlawful for an operator of a covered platform to condition access to or preferred
placement on the platform on the purchase or use of other products offered by the platform
operator that are not “part of or intrinsic to” the platform.109
Section 3(a)(5) resembles existing tying doctrine, with certain key differences. Under current law,
a plaintiff can prevail on a tying claim by showing that:
1. The defendant offered two distinct products;
2. The defendant conditioned the sale of one product (the tying product) on the
purchase of the other product (the tied product);
3. The defendant possessed sufficient economic power in the tying product market
to coerce purchasers into acceptance of the tied product; and
4. The defendant’s conduct affected a non-trivial amount of interstate commerce.110
Some courts have also required plaintiffs to demonstrate that a tying arrangement had
anticompetitive effects in the tied product market.111
104
See Fiona M. Scott Morton, et al.,
Equitable Interoperability: The “Super Tool” of Digital Platform Governance,
YALE T OBIN CTR. FOR ECON. POLICY, Discussion Paper No. 4 at 26 (July 13, 2021).
105 ABA COMMENTS,
supra note 56, at 14.
106
See, e.g., Howard A. Shelanski,
Unilateral Refusals to Deal in Intellectual and Other Property, 76 ANTITRUST L.J.
369, 380 (2009).
107
See, e.g., Krisztian Katona,
AICOA’s Data Security, Privacy, and Content Moderation Issues Call for Risk
Assessm ent, DISRUPTIVE COMPETITION PROJECT (June 7, 2022), https://www.project -disco.org/privacy/060722-aicoas-
data-security-privacy-and-content-moderation-issues-call-for-risk-assessment/; Josh Withrow,
The Revised Am erican
Innovation and Choice Online Act (AICOA) is Still Fundam entally Flawed , R ST. INST. (May 26, 2022),
https://www.rstreet.org/2022/05/26/revised-antitrust -bill-remains-fundamentally-flawed/.
108
See “Affirmative Defenses”
infra.
109 S. 2992, 117th Cong. § 3(a)(5) (2022) (Reported Version); May 25 Draft § 3(a)(5) (2022).
110 HOVENKAMP,
supra note 7, at 435 (summarizing the test employed by most federal circuit courts of appeals).
111
Id. at 435-36.
Congressional Research Service
16
link to page 5 link to page 23
The American Innovation and Choice Online Act (S. 2992)
Unlike this test, Section 3(a)(5) does not contain an explicit market-power requirement. As
discussed, the AICOA instead uses certain quantitative criteria and a “critical trading partner”
standard to identify the platforms that would be subject to its prohibitions.112
Section 3(a)(5) would also depart from the tying test employed by some federal courts of appeals
that requires proof of anticompetitive effects.113 Instead of requiring regulators to prove such
effects, S. 2992 would al ow defendants to rebut a
prima facie case under Section 3(a)(5) by
establishing an absence of competitive harm.114
Section 3(a)(5) has attracted special attention because of its possible impact on Amazon Prime—a
subscription service that offers Amazon customers fast shipping of eligible products, among other
benefits.115 An Amazon executive has argued that Section 3(a)(5) would prohibit the firm from
requiring third-party merchants to use Amazon’s fulfil ment services as a condition of
participating in Prime.116 According to the company, such a rule would harm consumers because
it would prevent Amazon from being able to guarantee prompt delivery of Prime packages.117
The AICOA contains a provision that appears to be directed at this concern. The bil contains a
rule of construction that insulates firms from liability triggered “solely” by offering a
“fee-for-service subscription that provides benefits to covered platform users.”118
Section 3(a)(5)’s effect on Amazon Prime and other services likely to fal under the act’s
coverage may ultimately depend on the meaning of the term “solely” in this rule of construction.
For example, while the rule might insulate Amazon from certain types of claims involving Prime,
this limiting language may preserve the possibility of liability for specific ways that Amazon
structures the Prime program. In such cases, Amazon may need to rely on the general affirmative
defenses discussed below.119
Section 3(a)(6): Use of Data
Section 3(a)(6) of S. 2992 would prohibit operators of covered platforms from using non-public
data generated by business users or the customers of business users to support their own
competing products or services.120
112
See “Covered Platforms”
supra.
113
See, e.g., Kaufman v. T ime Warner, 836 F.3d 137, 141 (2d Cir. 2016); Amey, Inc. v. Gulf Abstract & T itle Inc., 758
F.2d 1486, 1503 (11th Cir. 1985); Driskill v. Dallas Cowboys Football Club, Inc., 498 F.2d 321, 323 (5th Cir. 1974).
114 S. 2992, 117th Cong. § 3(b)(2)(A) (Reported Version); May 25 Draft § 3(b)(2).
115 Compare Fact Sheet: How the American Innovation and Choice Online Act (S. 2992) Would Break Amazon Prime
and Reduce Access to Amazon Basics, CHAMBER OF PROGRESS (Nov. 29, 2021), https://progresschamber.org/fact -
sheet -how-the-american-innovation-and-choice-online-act-s-2992-would-break-amazon-prime-and-block-amazon-
basics/ (arguing that S. 2992’s tying prohibition would eliminate the logistics model that makes Amazon Prime
possible), with Sen. Chuck Grassley, Q&A: Reining in Unbridled Big T ech (June 10, 2022),
https://www.grassley.senate.gov/news/news-releases/qanda-reining-in-unbridle d-big-tech (arguing that S. 2992
“explicitly does not impact subscription services, such as Amazon Prime”).
116 Brian Huseman,
Antitrust Legislation and the Unintended Negative Consequences for American Consumers and
Sm all Businesses, AMAZON (June 1, 2022), https://www.aboutamazon.com/news/policy-news-views/antitrust -
legislation-and-the-unintended-negative-consequences-for-american-consumers-and-small-businesses.
117
See id.
118 S. 2992, 117th Cong. § 3(c)(7)(A)(vi)(II) (Reported Version); May 25 Draft § 3(c)(8)(A)(vi)(II).
119
See “Affirmative Defenses”
infra.
120 S. 2992, 117th Cong. § 3(a)(6) (Reported Version); May 25 Draft § 3(a)(6).
Congressional Research Service
17
The American Innovation and Choice Online Act (S. 2992)
The provision appears to respond to concerns that tech platforms have used their unique access to
user data to identify and imitate popular offerings. For example, in October 2020, a House
subcommittee concluded that Amazon had used third-party sel er data to find profitable
opportunities to develop its own private-label products.121 Apple has also al egedly used
information from app developers to build competing offerings and integrate certain functionalities
into its iOS operating system.122
Chal enges to such practices have traditional y sounded in intel ectual-property (IP) law rather
than antitrust, to the extent that they involve possible patent infringement.123 The two bodies of
law stand in some tension: while antitrust safeguards competition, IP law offers temporary
monopolies over protected technology to incentivize innovation.124
Some commentators have argued that the innovation concerns driving IP law may extend to
certain
non-patented technology that tech platforms stand accused of copying.125 For example,
when a platform copies non-patented technology that nevertheless requires significant investment
in innovation, the platform’s conduct arguably raises the same free-rider problems that motivate
IP law.126
In addition to these innovation issues, anti-copying measures like Section 3(a)(6) may be
motivated by fairness considerations.127
In practice, however, these animating principles would not necessarily play a direct role in
Section 3(a)(6)’s application. S. 2992 would al ow a defendant to rebut a
prima facie case under
that provision by proving an absence of harm to
competition—not innovation or fairness.128 As
noted, those goals can pull in different directions.
Depending on the interpretation of the relevant language, the burden to prove an absence of
competitive harm may be fairly easy to satisfy. A platform’s entry into a new market—whether it
involves copying a rival product or not—wil typical y
increase the type of competition that
antitrust protects.129 To the extent that Section 3(a)(6) is motivated by concerns other than
121 INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 274-82.
122
See, e.g.,
id. at 361-64; Reed Albergotti,
How Apple Uses its App Store to Copy the Best Ideas, WASH. POST (Sept.
5, 2019), https://www.washingtonpost.com/technology/2019/09/05/how-apple-uses-its-app-store-copy-best-ideas/.
123 Herbert J. Hovenkamp,
Monopolizing and the Sherman Act, Faculty Scholarship at Penn Law 52 (Jan. 2022),
https://scholarship.law.upenn.edu/cgi/viewcontent.cgi?article=3772&context=faculty_scholarship .
124 Herbert Hovenkamp,
The Intellectual Property-Antitrust Interface,
in 3 ISSUES IN COMPETITION LAW AND POLICY
1979, 1979 (2008). Many commentators and judicial decisions have discussed the different strategies of IP law and
antitrust for promoting economic welfare.
See, e.g., SCM Corp. v. Xerox Corp., 645 F.2d 1195, 1203 ( 2d Cir. 1981)
(“T he conflict between the antitrust and patent laws arises in the methods they embrace that were designed to achieve
reciprocal goals. While the antitrust laws proscribe unreasonable restraints of competition, the patent laws reward the
inventor with a temporary monopoly that insulates him from competitive exploitation of his patented art.”).
125
See Sam Bowman,
Amazon’s Tightrope: Balancing Innovation and Competition on Amazon’s Marketplace, T RUTH
ON THE MARKET (Apr. 27, 2020), https://t ruthonthemarket.com/2020/04/27/amazons-tight rope-balancing-innovation-
and-competition-on-amazons-marketplace/.
126
See id.
127
See generally Francesco Ducci & Michael T rebilcock,
The Revival of Fairness Discourse in Competition Policy, 64
ANTITRUST BULLETIN 79 (2019).
128 S. 2992, 117th Cong. § 3(b)(2)(A) (2022) (Reported Version); May 25 Draft § 3(b)(2) (2022).
129
See Hovenkamp,
supra note 124, at 1979 (“As a general proposition, the more firms that offer a product the more
competitive will be its output and price.”).
Congressional Research Service
18
The American Innovation and Choice Online Act (S. 2992)
competition, then, the affirmative defense for an absence of competitive harm may narrow the
provision’s scope in ways that are inconsistent with its theoretical underpinnings.
Section 3(a)(8): App Preinstallation and Steering
Section 3(a)(8) of the AICOA would make it unlawful for the operator of a covered platform to
“material y restrict or impede” platform users from uninstal ing preinstal ed apps or changing
default settings that steer users to the platform operator’s products or services.130 Such restrictions
would be permissible, however, when necessary for the security or functioning of the platform or
to prevent data from being transferred to the government of the People’s Republic of China or a
foreign adversary.131
The provision is a response to concerns that Apple and Google—which control the leading mobile
operating systems, iOS and Android—have used app preinstal ation, default settings, and related
contractual restrictions to favor their own apps over rivals.132
Section 3(a)(9): Self-Preferencing Involving Platform Interfaces
Section 3(a)(9) of the May 25 draft would prohibit operators of covered platforms from using
their platform interfaces (including search or ranking functionalities) to treat their own products
or services more favorably than those of other business users “in a manner that is inconsistent
with the neutral, fair, and non-discriminatory treatment of al business users.”133
The provision appears to overlap with Section 3(a)(1)’s self-preferencing prohibition, but there
are three apparent differences.
First, unlike Section 3(a)(1), Section 3(a)(9) would not require regulators to
establish competitive harm.
Second, Section 3(a)(9)’s prohibition is limited to a platform operator’s use of its
platform interface, while Section 3(a)(1) is not.
Third, Section 3(a)(9) would require regulators to establish that a platform
operator’s conduct was “inconsistent with the neutral, fair, and
non-discriminatory treatment of al business users,” while Section 3(a)(1) does
not include such a requirement.134
Section 3(a)(10): Retaliation for Reports to Law Enforcement
Section 3(a)(10) of the AICOA would make it unlawful for the operator of a covered platform to
retaliate against business users that raise good-faith concerns with law-enforcement authorities
about actual or potential violations of law.135
130 S. 2992, 117th Cong. § 3(a)(8) (Reported Version); May 25 Draft § 3(a)(8).
131 S. 2992, 117th Cong. § 3(a)(8)(A)-(B) (Reported Version); May 25 Draft § 3(a)(8)(A)-(B).
132
See INVESTIGATION OF COMPETITION IN DIGITAL MARKETS,
supra note 1, at 352-58, 379.
133 May 25 Draft § 3(a)(9). T he reported version of the bill would prohibit operators of covered platforms from using
their platform interfaces to favor their products or services “ relative to those of another business user than under [sic]
standards mandating the neutral, fair, and nondiscriminatory treatment of all business users.” S. 2992, 117th
Cong. § 3(a)(9) (Reported Version).
134 May 25 Draft § 3(a)(1), (a)(9).
135 S. 2992, 117th Cong. § 3(a)(10) (Reported Version); May 25 Draft § 3(a)(10).
Congressional Research Service
19
The American Innovation and Choice Online Act (S. 2992)
Comparing Different Versions of the AICOA: S. 2992 and H.R. 3816
While S. 2992 would bar operators of covered platforms from engaging in 10 categories of conduct, H.R. 3816
contains 13 prohibitions.
See S. 2992, 117th Cong. § 3(a)(1)-(10) (2022) (Reported Version); May 25
Draft § 3(a)(1)-(10) (2022); H.R. 3816, 117th Cong. § 2(a)-(b) (2021). The additional offenses in the House
committee-reported bil involve restrictions on business users’ communications with platform users, interference
with business users’ pricing of their products and services, and specific types of interoperability.
See H.R. 3816,
117th Cong. § 2(b)(6), (b)(8)-(9).
Although several of the other prohibitions in S. 2992 and H.R. 3816 overlap, there are also key differences. Unlike
S. 2992—which would make harm to competition an element of three offenses—none of H.R. 3816’s prohibitions
would require proof of competitive harm as part of a plaintiff’s case-in-chief. S. 2992, 117th Cong. § 3(a)(1)-(3)
(Reported Version); May 25 Draft § 3(a)(1)-(3); H.R. 3816, 117th Cong. § 2(a)-(b).
Affirmative Defenses
Section 3(b) of the bil would provide operators of covered platforms with several affirmative
defenses.
Section 3(b)(1) of the May 25 draft would offer an affirmative defense to al of the bil ’s
prohibitions for conduct that is “reasonably tailored and reasonably necessary, such that the
conduct could not be achieved through material y less discriminatory means, to”:
Prevent a violation of, or comply with, federal or state law;
Protect safety, user privacy, the security of non-public data, or the security of a
covered platform; or
Maintain or substantial y enhance the “core functionality” of a covered
platform.136
As discussed, Section 3(b)(2) would offer an affirmative defense to the prohibitions in
Sections 3(a)(4)-(10) for conduct that “has not resulted in and would not result in material harm
to competition.”137
Defendants would bear the burden of proving these defenses by a preponderance of the evidence
(
i.e., by proving that the relevant propositions are more likely true than not true).138
The bil ’s proponents have argued that the defenses appropriately place the burden to defend
potential y anticompetitive conduct on platform operators, who have more information about their
products than regulators.139 The May 25 version of the legislation also relaxed certain language in
136 May 25 Draft § 3(b)(1). T he parallel defense in the reported version of the bill would allow defendants to rebut a
prim a facie case under Sections 3(a)(1)-(3) by showing that their conduct was “ narrowly tailored, nonpretextual, and
reasonably necessary” to achieve any of the goals identified in the May 25 version. S. 2992, 117th Cong. § 3(b)(1)
(Reported Version). For the other seven offenses, the reported version of the bill would allow defendants to re but a
prim a facie case by showing that their conduct “ was narrowly tailored, could not be achieved through less
discriminatory means, was nonpretextual, and was reasonably necessary” to achieve any of those same goals. S. 2992,
117th Cong. § 3(b)(2)(B) (Reported Version).
137 S. 2992, 117th Cong. § 3(b)(2)(A) (Reported Version); May 25 Draft § 3(b)(2).
138
See S. 2992, 117th Cong. § 3(b)(1)-(2) (Reported Version); May 25 Draft § 3(b)(4); United States v. Watkins, 10
F.4th 1179, 1184-85 (11th Cir. 2021).
139 Scott Morton, et al.,
supra note 9, at 3-4.
Congressional Research Service
20
The American Innovation and Choice Online Act (S. 2992)
previous iterations of Section 3(b)(1) in response to criticism that the earlier defenses would have
been overly difficult for defendants to establish.140
Critics have contended that even the less demanding language in the May 25 draft may chil
platforms’ efforts to promote privacy and data security. In particular, these commentators have
argued that platforms often engage in duplicative efforts to protect customers, which may leave
them vulnerable to claims that they could achieve their goals through “material y less
discriminatory means.”141
Comparing Different Versions of the AICOA: S. 2992 and H.R. 3816
The affirmative defenses in H.R. 3816 differ from those in S. 2992. As discussed, S. 2992 would al ow defendants
to rebut a
prima facie case under Sections 3(a)(4)-(10) by proving an absence of “material harm to competition” by
a preponderance of the evidence.
See S. 2992, 117th Cong. § 3(b)(2)(A) (2022) (Reported Version); May 25
Draft § 3(b)(2) (2022). In contrast, the corresponding defense in H.R. 3816 would require defendants to prove an
absence of harm to “the competitive process” by clear and convincing evidence. H.R. 3816, 117th Cong. § 2(c)(1)
(2021).
The clear-and-convincing-evidence standard is more stringent than a preponderance-of-the-evidence burden. The
latter requires evidence that makes it more likely than not that a proposition is true.
See, e.g., United States v.
Watkins, 10 F.4th 1179, 1184-85 (11th Cir. 2021). By contrast, the clear-and-convincing-evidence test demands
that the evidence makes a contention “highly probable.” Colorado v. New Mexico, 467 U.S. 310, 316 (1984).
According to a large empirical study, many judges regard 75% probability as a reasonable approximation of the
clear-and-convincing-evidence standard.
See C.M.A. McCauliff,
Burdens of Proof: Degrees of Belief, Quanta of Evidence,
or Constitutional Guarantees?, 35 VAND. L. REV. 1293, 1328-29 (1982).
The House committee-reported bil would also offer affirmative defenses related to privacy and data security.
Defendants could rebut a
prima facie case under H.R. 3816 by establishing by clear and convincing evidence that
their conduct was “narrowly tailored, could not be achieved through less discriminatory means, was
nonpretextual, and was necessary” to prevent a violation of law, protect user privacy, or protect non-public data.
H.R. 3816, 117th Cong. § 2(c)(2). While the evidentiary standards differ, the substantive language in this defense is
similar to—but not identical with—paral el affirmative defenses in the reported version of S. 2992.
See note 136
supra. As discussed, Senator Klobuchar’s May 25 draft relaxed the relevant language in several respects.
Unlike both bil s in the Senate, H.R. 3816 would not offer an affirmative defense related to the maintenance or
enhancement of platform functionality. The House committee-reported bil would, however, offer a defense that
does not appear in either version of S. 2992. Specifical y, defendants could rebut a
prima facie case under H.R.
3816 by establishing by clear and convincing evidence that their conduct “increases consumer welfare.” H.R. 3816,
117th Cong. § 2(c)(3).
Enforcement
S. 2992 would grant enforcement authority to the DOJ, the FTC, and state attorneys general.142
The bil does not contain a private right of action.
140
See, e.g., Letter from Aurelian Portuese to Sen. Dick Durbin, et al., 3 (Jan. 19, 2022), https://www2.itif.org/2022-
IT IF-letter-S2992.pdf (criticizing an earlier iteration of Section 3(b) for establishing “ insurmountable thresholds” for
the relevant affirmative defenses); Letter from T imothy Powderly to Sen. Dick Durbin, et al., 2 (Jan. 18, 2022),
https://9to5mac.com/wp-content/uploads/sites/6/2022/01/Apple-letter-full.pdf (criticizing an earlier iteration of
Section 3(b) for establishing “ a nearly insurmountable test”). For the relevant language in the reported version of the
bill, see note 136
supra.
141
See, e.g., Lawrence J. Spiwak,
The Third Time is Not the Charm: Significant Problems Remain With Senator
Klobuchar’s Antitrust Reform Bill, T HE FEDERALIST SOC’Y (June 7, 2022), https://fedsoc.org/commentary/fedsoc-
blog/the-third-time-is-not-the-charm-significant -problems-remain-with-senator-klobuchar-s-antitrust -reform-bill.
142 S. 2992, 117th Cong. § 3(c)(1) (Reported Version); May 25 Draft § 3(c)(1).
Congressional Research Service
21
The American Innovation and Choice Online Act (S. 2992)
As discussed, the legislation would empower the DOJ and FTC to jointly designate firms that
meet the relevant criteria as covered platforms.143 Such designations would be valid for seven
years, though the DOJ and FTC would be al owed to reevaluate designation decisions upon
receiving a request showing that a platform no longer meets the relevant criteria.144 Designated
platforms would also be permitted to seek judicial review of their designations.145
Section 4 of the bil would direct the DOJ and FTC to jointly issue guidelines outlining their
interpretation of the “material y harm competition” standard in Section 3(a) and the affirmative
defenses in Section 3(b), in addition to their policies regarding civil penalties.146
Under the May 25 version of the legislation, firms that violate the bil ’s prohibitions would be
liable for up to 10% of their total U.S. revenue for the period in which the violation occurred.147
In cases of recurring violations, the legislation would authorize courts to order a firm’s chief
executive officer to forfeit any compensation received during the 12 months preceding the filing
of a complaint.148
Comparing Different Versions of the AICOA: S. 2992 and H.R. 3816
The versions of the AICOA pending in the Senate and the House would adopt different enforcement schemes. As
discussed, the May 25 version of S. 2992 would authorize penalties of up to 10% of a defendant’s total U.S.
revenue over the course of a violation. May 25 Draft § 3(c)(6)(B) (2022). Under H.R. 3816, by contrast, violators
would face penalties of up to (1) 15% of their total U.S. revenue for the previous calendar year, or (2) 30% of the
U.S. revenue of entities “affected or targeted” by the offending conduct, calculated over the course of a violation.
H.R. 3816, 117th Cong. § 2(f)(1) (2021).
H.R. 3816 also explicitly contemplates that divestiture orders may be appropriate remedies for violations of the
bil . The legislation provides that, if a court determines that a violation arises from a “conflict of interest” related
to a platform operator’s control of multiple business lines, the court “shal consider” and “may order” divestiture
of the business lines giving rise to the conflict. H.R. 3816, 117th Cong. § 2(f)(2)(D). S. 2992 does not contain an
analogous provision, but instead makes clear that it does not “prevent or limit” regulators from seeking equitable
relief. S. 2992, 117th Cong. § 3(c)(5)(C)(i )(V) (Reported Version); May 25 Draft § 3(c)(6)(C)(i )(V).
One of the most significant differences involves the legislation’s would-be enforcers: H.R. 3816 includes a private
right of action for treble damages, while S. 2992 does not. H.R. 3816, 117th Cong. § 6(a).
Conclusion
The AICOA reflects many of the concerns about digital competition that have occupied
congressional attention over the past several years. It also implicates difficult questions involving
innovation, privacy, data security, and online speech. Regardless of whether S. 2992 or H.R. 3816
ultimately becomes law, the issues motivating the bills may continue to garner legislative interest.
143 S. 2992, 117th Cong. § 3(d)(1) (Reported Version); May 25 Draft § 3(d)(1).
144 S. 2992, 117th Cong. § 3(d)(1)(C), (d)(2) (Reported Version); May 25 Draft § 3(d)(1)(C), (d)(2).
145 S. 2992, 117th Cong. § 3(d)(3) (Reported Version); May 25 Draft § 3(d)(3).
146 S. 2992, 117th Cong. § 4(a) (Reported Version); May 25 Draft § 4(a).
147 May 25 Draft § 3(c)(6)(B). Under the reported version of S. 2992, violators would be liable for up to 15 percent of
their total U.S. revenue for the period in which the violation occurred. S. 2992, 117th Cong. § 3(c)(5)(B) (Reported
Version).
148 S. 2992, 117th Cong. § 3(c)(5)(D) (Reported Version); May 25 Draft § 3(c)(6)(D).
Congressional Research Service
22
The American Innovation and Choice Online Act (S. 2992)
Author Information
Jay B. Sykes
Legislative Attorney
Disclaimer
This document was prepared by the Congressional Research Service (CRS). CRS serves as nonpartisan
shared staff to congressional committees and Members of Congress. It operates solely at the behest of and
under the direction of Congress. Information in a CRS Report should n ot be relied upon for purposes other
than public understanding of information that has been provided by CRS to Members of Congress in
connection with CRS’s institutional role. CRS Reports, as a work of the United States Government, are not
subject to copyright protection in the United States. Any CRS Report may be reproduced and distributed in
its entirety without permission from CRS. However, as a CRS Report may include copyrighted images or
material from a third party, you may need to obtain the permission of the copyright holder if you wish to
copy or otherwise use copyrighted material.
Congressional Research Service
R47228
· VERSION 5 · UPDATED
23