The Algorithmic Cartel

Modernising Antitrust for the Era of Automated Coordination

In November 2025, the United States Department of Justice announced a proposed consent judgment with RealPage, Inc., a software company whose revenue management tools operated across residential housing markets by aggregating tenant and rental data from thousands of competing property managers.

The proposed consent judgment is one of the clearest federal templates yet for regulating the architecture of pricing software. It treats algorithmic pricing tools not merely as passive decision-support systems, but as potential instruments for unlawful information sharing and price alignment, advancing the enforcement theory that software can function as an instrument of unlawful price coordination even where no traditional human “meeting of minds” is pleaded in conventional cartel terms.

This marks a regulatory turning point in what commentators increasingly describe as the “algorithmic cartel” and what enforcers more often frame as algorithmic coordination, information sharing and pricing alignment - a phenomenon in which automated systems achieve convergent, supra-competitive market outcomes by processing competitively sensitive data in real-time without explicit human agreement.1 What makes this development urgent is not the novelty of price-fixing itself, but the speed and opacity at which coordination now occurs. A landlord accepting a pricing recommendation may not realise that the system is being treated by enforcers as part of an alleged anticompetitive information-sharing architecture. An autonomous system designed to maximise profit may independently learn that coordination produces better returns than competition. These scenarios challenge the foundational legal concept that cartels require a “meeting of minds.” The data-hub version of the problem is no longer theoretical; the fully autonomous version is now close enough to the market to worry regulators.

The RealPage Proposed Consent Judgment: Architectural Restrictions on Algorithmic Influence

The proposed DOJ-RealPage settlement addresses coordination through three distinct mechanisms: the age of competitor data used to train pricing models, the geographic scope of that data and the architecture of software choice itself. At the centre of the DOJ’s complaint was RealPage’s YieldStar and AI Revenue Management (AIRM) platforms, which functioned as a “melting pot” of confidential competitor information.2,3 By pooling granular, non-public lease data from tens of thousands of properties (including actual transactional prices, future occupancy levels and specific lease terms), RealPage’s models generated pricing recommendations that maximised total industry revenue rather than reflecting the independent competitive interests of individual landlords.4,5

The innovation of the settlement lies in its distinct treatment of runtime operation versus model training. The DOJ recognised that the most immediate collusion risk occurs when an algorithm uses a competitor’s real-time, non-public data to adjust prices in the current market.6 The proposed judgment would prohibit RealPage from utilising any competitively sensitive information (CSI) from unaffiliated properties during runtime operations of its revenue management tools.2,6 This restriction would ensure that a landlord’s daily pricing updates are based either on their own proprietary data or on information that is “readily accessible to the general public”.6

Regulators took a different approach toward model training, acknowledging that AI systems require access to historical datasets to maintain predictive accuracy and improve market efficiency.2,7 The proposed judgment would permit RealPage to train its models on non-public competitor data, but only if that data is at least 12 months old.2,6 This temporal buffer neutralises the strategic potency of the data while preserving its utility for identifying long-term cyclical trends in supply and demand.

The proposed judgment would also bar RealPage from using models that determine geographic effects narrower than state level.2,6 That requirement is broader than the rental markets alleged in the complaint and is designed to reduce the risk of hyper-local price alignment.8,9 By preventing the algorithm from learning and enforcing localised pricing patterns, the proposed judgment aims to dilute the risk of tacit coordination within a specific neighbourhood or census block while still allowing the software to account for regional economic shifts.

Choice Architecture and the Coercion of Compliance

The DOJ’s investigation revealed that the effectiveness of the RealPage system relied not just on the data itself, but on the choice architecture designed to suppress independent decision-making by property managers. The complaint alleged that RealPage’s software was engineered to make accepting price recommendations a frictionless process while making overrides difficult and time-consuming.3,5 Property managers could “bulk accept” multiple recommendations but were required to provide “specific business commentary” for every rejection.3,5 These justifications were reviewed by RealPage “pricing advisors” and could be escalated to regional managers. This created a psychological and bureaucratic deterrent against competitive pricing.3,5

The proposed consent judgment would require RealPage to eliminate these coercive design features. Software settings, such as “Auto-Accept” or “Governor” features (which were alleged to favour price increases over decreases), would need to be symmetrical and require active, manual configuration by the user.6 This provision establishes a core principle emerging in modern antitrust enforcement: algorithmic tools must serve as aids to independent human judgment rather than as automated mandates for market alignment.10,11

The Agri Stats Precedent: Benchmarking as a Coordination Hub

While the RealPage case focused on the technology of AI itself, the Agri Stats litigation demonstrates how traditional data-benchmarking services have evolved into instruments of algorithmic coordination. Agri Stats sits alongside RealPage as a broader information-exchange precedent, although the procedural posture differs. Agri Stats collected and distributed detailed financial and production data among the leading processors of chicken, pork and turkey.12,13 The DOJ’s case alleges that Agri Stats’ reports enabled meat processors to use competitively sensitive information to stabilise or increase prices and reduce output across the US protein industry.12,14

Separately, in the private wage-suppression litigation, Agri Stats settlements required changes to labour-related reporting, including removal or redaction of plant-level labour data, preventing meatpackers from monitoring the specific compensation strategies of their rivals.13,15 This redact-at-source requirement targets the transparency problem that facilitates coordination in oligopolistic markets.

The Withdrawal of Regulatory Safe Harbours

The Agri Stats and RealPage cases must be viewed against a broader regulatory shift: the withdrawal of the 1996 and 2000 antitrust safety zones for information sharing.13,16 For decades, firms operated under the assumption that sharing data was legal if it was managed by a third party, involved at least five participants, and was at least three months old.2,13 The withdrawal began earlier than the current algorithmic-pricing cases. DOJ withdrew the health-care information-sharing policy statements in February 2023, the FTC followed in July 2023, and both agencies withdrew the 2000 Collaboration Guidelines in December 2024.7,16 The rationale was consistent: existing safe harbours were overly permissive and failed to account for the ability of modern algorithms to de-anonymise aggregated data and exploit even three-month-old information in real-time.

The Strategic Uncertainty Crisis in Hospitality: The CMA Hotel Investigation

On 24 February 2026, announced publicly on 2 March 2026, the United Kingdom’s Competition and Markets Authority (CMA) launched a Competition Act 1998 investigation into three global hotel giants (Hilton, Marriott and IHG) alongside the data analytics provider STR, owned by CoStar Group.17,18 The investigation centres on the suspected sharing of competitively sensitive information through STR’s benchmarking tools.17,19 STR monitors over 90,000 hotels globally and provides the industry with reports on occupancy, average daily rates (ADR), and revenue per available room (RevPAR).18,20

The CMA’s theory of harm focuses on reduction of strategic uncertainty. In a competitive market, a firm’s inability to predict its rival’s pricing and capacity decisions drives it to lower prices or improve service.17,20 The CMA is investigating whether the use of a common data-services provider reduced strategic uncertainty between competing hotel chains by enabling the exchange of competitively sensitive information. The CMA has not reached a view on whether there is sufficient evidence of infringement.17 This case matters because it demonstrates how traditional benchmarking platforms may transition into hubs for algorithmic coordination in highly dynamic service markets.

Agentic AI and the Risk of Autonomous Collusion

The hotel investigation sits against a broader CMA concern about pricing algorithms, AI systems and the possibility that automated tools may facilitate collusion or reduce strategic uncertainty.21 The CMA warned that when competing firms deploy agents programmed to maximise profit, these agents may independently discover that coordination produces returns superior to aggressive competition.15,21 This autonomous collusion presents a major enforcement challenge because it can occur without any direct communication between the human owners of the agents.15,21

Transatlantic Divergence: The Doctrinal Chasm

The central legal debate in both US and EU jurisdictions is whether independent adoption of the same pricing algorithm constitutes a violation of antitrust laws. This debate exposes a growing gap between established law and the realities of digital markets.1,2

US Jurisprudence: The “Agreement” Requirement vs. Plus Factors

Under Section 1 of the Sherman Act, the US government must prove the existence of a “contract, combination, or conspiracy”.10 In the Gibson v. Cendyn case (August 2025), the Ninth Circuit affirmed the dismissal of a class action against several hotel chains, ruling that simply subscribing to the same pricing software was insufficient to establish an agreement.8 The court described this as “consciously parallel conduct,” which is not illegal unless accompanied by “plus factors” that suggest a meeting of minds.8

Other US courts have adopted a broader view. In Duffy v. Yardi (December 2024), a district court in Washington denied a motion to dismiss, finding that the software provider’s marketing (which promised to help landlords “maximize rents across the industry”) functioned as an “invitation to conspire”.8,12 By accepting this invitation and providing their commercially sensitive data to the melting pot, the landlords effectively entered into an unlawful horizontal agreement.8,12

EU Framework: The “Concerted Practice” and Article 101 TFEU

Article 101(1) TFEU prohibits not only agreements but also “concerted practices” (a broader category that encompasses any direct or indirect contact between competitors intended to influence market behaviour).2,22 This framework is inherently more adaptable to algorithmic coordination. Linsey McCallum, the Deputy Director General of the European Commission’s competition arm, stated in July 2025 that the Commission had identified red flags in multiple confidential investigations where algorithms facilitate or monitor coordination between competitors.15,23

In the EU, two scenarios are at the forefront of the debate. The “Predictable Agent” refers to algorithms that respond to market signals in a way that leads to tacit coordination. The “Digital Eye” describes autonomous systems that independently learn that collusion is the most profitable strategy.1,2 European enforcers are moving toward an attribution-and-control model: undertakings cannot avoid liability merely because pricing decisions are mediated through software that they chose, configured, supplied with data or failed to supervise.24

Sectoral Enforcement in the European Economic Area

While the European Commission pursues large-scale investigations, national competition authorities are taking the lead in specific sectors. These investigations provide practical insights into how Article 101 is being applied to algorithmic cartels.24

Poland’s UOKiK: Banking and Pharmaceuticals

In September 2025, the President of the Polish Office for Competition and Consumer Protection (UOKiK) confirmed two major investigations into algorithmic pricing. The banking investigation reportedly involves several Polish banks that allegedly utilised algorithms fed by data from the country’s largest credit risk database and their own non-public internal information to coordinate the pricing of consumer loans and mortgages.15 The pharmaceutical wholesale probe reportedly concerns three major wholesalers controlling 80% of the market, who allegedly used IT systems to exchange commercially sensitive information on drug margins, prices, and volumes sold through affiliated pharmacies.12,25 This should be distinguished from earlier UOKiK proceedings concerning pharmaceutical wholesalers and software providers, which also involved alleged exchanges of commercially sensitive information but arose in a separate procedural context.25

These cases matter because they target both the spokes (the banks and wholesalers) and the hub (the software providers), reflecting a regulatory consensus that the facilitators of algorithmic collusion share in the liability.26

Legislative Frontiers: Bridging the Proof Gap

The difficulty of proving a meeting of minds in the era of deep learning has prompted legislative action. These bills seek to change the presumptions of antitrust law, shifting the burden of proof from the regulator to the firm.

The Preventing Algorithmic Collusion Act was reintroduced in January 2025 by Senator Amy Klobuchar as Senate Bill 232. This legislation aimed to modernise the Sherman Act by creating a legal presumption of collusion for certain uses of pricing algorithms.2 The bill would have prohibited the use of algorithms that rely on non-public competitor data and mandated algorithmic audits.2,10 The bill was introduced on 23 January 2025 and referred to the Senate Judiciary Committee. As at publication, it had not advanced beyond that stage.27 Industry objections have focused on innovation, overbreadth and tension with conventional rule-of-reason analysis.2

While federal algorithmic-collusion legislation has not advanced, California and New York emerged as the primary antitrust frontier in 2025 and 2026.28,29 California AB 325, approved in October 2025 and operative from 1 January 2026, amends the Cartwright Act to prohibit the use or distribution of a common pricing algorithm as part of an unlawful contract, combination or conspiracy in restraint of trade. It also prohibits using or distributing such an algorithm where a person coerces another to adopt a recommended price or commercial term, and lowers the pleading threshold by requiring only plausible allegations of conspiracy rather than facts excluding independent action.28,30 AB 325 will nevertheless make due diligence practically essential, because businesses will need to understand whether their tools use competitor data and whether any feature could be characterised as coercing adoption of recommended prices.

New York moved on two tracks: a housing-specific restriction on algorithmic rent-setting and a separate consumer-facing disclosure law for algorithmic pricing based on personal data. AB A1417B bans landlords from using rent-setting software that utilises non-public competitor data.15,28 The separate Algorithmic Pricing Disclosure Act, now in effect, requires most businesses using personalised algorithmic pricing based on consumer personal data to display a clear disclosure near the price.28

Conceptual Models of Algorithmic Collusion

To effectively regulate algorithmic pricing, authorities have identified three distinct scenarios of harm, each presenting unique evidentiary challenges.

The first scenario is the Automated Traditional Cartel. Here the algorithm is merely a tool used to implement an existing human agreement. Competitors reach a meeting of minds via email or meeting and then program their repricing software to enforce the arrangement.1 This occurred in the 2016 UK Posters case, where sellers on Amazon used software to ensure they never undercut each other.1,23 Existing Section 1/Article 101 doctrine easily captures this scenario because criminal intent remains human.

The second scenario is Hub-and-Spoke Coordination. Multiple competitors delegate their pricing decisions to the same common algorithm (the hub). Even if competitors do not communicate directly, they knowingly enter into a shared system that pools their sensitive data and provides aligned recommendations.1,31 The RealPage and Agri Stats cases exemplify this model. The legal challenge lies in proving that the spokes (competitors) understood the collaborative nature of the hub.15,26

The third scenario is the “Digital Eye.” AI systems trained to maximise profits independently discover that coordination produces better outcomes than competition.1 Through reinforcement learning, the software learns to signal its intention to maintain high prices and to punish any rival that cheats by lowering them.2,24 Because there is no human agreement and no shared hub, this scenario falls into a legal grey zone where traditional doctrines of coordinated conduct are difficult to apply.1,2

Economic Implications: The End of Strategic Uncertainty

Algorithmic pricing has altered the game theory of market competition. Economic theory suggests that price transparency benefits consumers. In digital markets, however, this transparency is often asymmetrical.1,23

Algorithms allow firms to monitor rivals with extreme speed and granularity. This high-frequency monitoring reduces reaction time to zero, meaning any firm attempting to gain market share by lowering prices is instantly matched.1,15 When price-cutting becomes futile, competition shifts from a race to the bottom to an equilibrium at supra-competitive levels. The allegations against RealPage’s Governor feature illustrate this price stickiness in the upward direction.3,6

The RealPage settlement also highlighted how software design can nudge users toward collusive outcomes. By creating an environment where diverging from a recommendation requires high administrative effort, software providers effectively strip human agents of their independent competitive instincts.3,5 Even if a firm has no intent to collude, its participation in a shared pricing ecosystem can lead to anti-competitive market effects through algorithmic coercion.11,28

The Future of Enforcement: From Hindsight to Oversight

The settlements of late 2025 and early 2026 indicate a permanent shift in how antitrust authorities view technology. The wait-and-see approach of the early 2020s has been replaced by aggressive, proactive enforcement.21,23

Enforcers now demand that companies conduct internal audits of their pricing tools. The CMA and EC have emphasised that businesses must understand the sources of underlying training data and ensure that their software’s objectives are not promoting aligned or elevated pricing.15,26 There is growing academic and regulatory push to apply the precautionary principle to algorithmic markets.34 This approach argues that because of unknown unknowns associated with deep learning and the speed at which digital markets tip into monopolies, regulators should intervene early to prevent harm rather than waiting for empirical proof of an antitrust violation.34 The CMA hotel case is not a market investigation; it is a Competition Act 1998, Chapter I investigation. But it sits within a broader UK move from passive observation toward earlier scrutiny of data hubs, pricing tools and algorithmic systems.15

Conclusion: Mapping the Unresolved Territory

The algorithmic cartel exposes fundamental gaps in 20th-century competition law. From human-mediated agreements to data-mediated convergence, the regulatory environment has shifted dramatically. The Sherman Act and Article 101 TFEU are being stretched in real-time to cover automated coordination. RealPage and Agri Stats demonstrate that regulators are finding ways to adapt legacy doctrines to algorithmic reality.

If entered, the RealPage model - defined by 12-month data ageing, statewide granularity restrictions and mandatory symmetry in software settings - is likely to become an influential template for algorithmic-pricing compliance.2,6 Investigations in the UK (hotels) and Poland (pharmaceuticals) signal that regulators will no longer tolerate third-party data hubs that eliminate strategic uncertainty.

Yet the law remains uncertain on the question that will define enforcement in the next phase. The Digital Eye scenario involves autonomous systems that collude without human direction or shared data through independent learning that coordination maximises profit - a possibility that falls into a regulatory grey zone. No existing doctrine cleanly captures it. The hard case is narrower but more profound: autonomous tacit coordination without contact, without a shared hub, without competitor data exchange and without an attributable instruction to collude. That is where the gap between economic harm and legal doctrine remains most exposed. Section 1 of the Sherman Act requires an agreement; Article 101 TFEU can capture indirect contact and concerted practices and liability can attach to conduct implemented through tools or agents. But an autonomous system that learns to collude through reinforcement learning presents none of these features. It may produce identical outcomes to a cartel, with identical consumer harm, but the legal tools designed to address cartels do not fit. This gap between harm and remedy is the unresolved question that will occupy enforcement agencies and courts throughout 2026 and beyond. For businesses, the path forward requires moving from algorithmic blind faith to architectural compliance - designing systems that compete rather than coordinate, using public data instead of competitor CSI, with responsibility for competitive behaviour now resting in the code itself.

References

1 Algorithmic Collusion: Corporate Accountability and the Application of Art. 101 TFEU, https://www.europeanpapers.eu/europeanforum/algorithmic-collusion-corporate-accountability-application-art-101-tfeu

2 Algorithmic Tacit Collusion: Addressing the Gaps in Article 101(1)(a) of the TFEU - Knight-Georgetown Institute, https://kgi.georgetown.edu/wp-content/uploads/2026/01/Algorithmic-Tacit-Collusion_Brambilla_17.pdf

3 Proposed DOJ settlement provides guidance on use of competitive information in algorithmic pricing tools - Hogan Lovells, https://www.hoganlovells.com/en/publications/proposed-doj-settlement-provides-guidance-on-use-of-competitive-information

4 United States of America et al. v. RealPage, Inc. et al.; Proposed Final Judgment and Competitive Impact Statement - Federal Register, https://www.federalregister.gov/documents/2025/12/05/2025-21966/united-states-of-america-et-al-v-realpage-inc-et-al-proposed-final-judgment-and-competitive-impact

5 Last Year’s Rent: RealPage Reaches Settlement Agreement with the Department of Justice in Algorithmic Pricing Case | Mintz, https://www.mintz.com/insights-center/viewpoints/2191/2025-12-01-last-years-rent-realpage-reaches-settlement-agreement

6 United States of America et al. v. RealPage, Inc. et al. Proposed Final Judgment and Competitive Impact Statement - Federal Register, https://www.federalregister.gov/documents/2026/01/21/2026-01009/united-states-of-america-et-al-v-realpage-inc-et-al-proposed-final-judgment-and-competitive-impact

7 Practical Takeaways From the DOJ’s Algorithmic Pricing Settlement, https://www.paulweiss.com/insights/client-memos/practical-takeaways-from-the-doj-s-algorithmic-pricing-settlement

8 Ninth Circuit Clarifies Antitrust Implications of Algorithmic Pricing, https://www.arnoldporter.com/en/perspectives/advisories/2025/08/antitrust-implications-of-algorithmic-pricing

9 DOJ Press Release: Justice Department Requires RealPage to End Sharing of Competitively Sensitive Information and Redesign Key Software Tools, 24 November 2025, https://www.justice.gov/opa/pr/justice-department-requires-realpage-end-sharing-competitively-sensitive-information-and

10 FTC and DOJ Statement of Interest, Cornish-Adebiyi v Caesars Entertainment, filed 28 March 2024; FTC press release, “FTC and DOJ File Statement of Interest in Hotel Room Algorithmic Price-Fixing Case”, 28 March 2024, https://www.justice.gov/archives/opa/pr/justice-department-and-federal-trade-commission-file-statement-interest-hotel-room

11 Corporate-Tech Landlordism REVISED 19Aug2025 - Stanford Law School, https://law.stanford.edu/wp-content/uploads/2025/07/Corporate-Tech-Landlordism-REVISED-19Aug2025.pdf

12 Gig Platforms as Hub-and-Spoke Arrangements and Algorithmic Pricing: A Comparative EU-US Antitrust Analysis - OpenEdition Books, https://books.openedition.org/putc/15512

13 Settlements Reached with Agri Stats in Broilers, Turkey, Pork, https://www.hbsslaw.com/press/pork-antitrust/settlements-reached-with-agri-stats-in-broilers-turkey-pork-antitrust-suits-over-price-fixing-allegations

14 Expect more turmoil and change for the 2026 ag sector - Investigate Midwest, https://investigatemidwest.org/2026/01/08/expect-more-turmoil-and-change-for-the-2026-ag-sector/

15 CMA Case Page: Suspected sharing of competitively sensitive information by Hilton, IHG, Marriott and STR (CoStar), CA98/01/2026, https://www.gov.uk/cma-cases/suspected-sharing-of-competitively-sensitive-information-by-hilton-ihg-marriott-and-str-costar

16 DOJ Press Release: Justice Department Withdraws Outdated Enforcement Policy Statements, February 2023, https://www.justice.gov/archives/opa/pr/justice-department-withdraws-outdated-enforcement-policy-statements; FTC withdrawal of healthcare antitrust policy statements, July 2023; DOJ/FTC joint withdrawal of 2000 Collaboration Guidelines, December 2024

17 UK Watchdog Probes Data-Sharing Among Hotel Giants - PYMNTS.com, https://www.pymnts.com/data/2026/uk-watchdog-probes-data-sharing-among-hotel-giants/

18 Hilton, IHG and Marriott under CMA investigation for information sharing - The Caterer, https://www.thecaterer.com/news/hilton-ihg-and-marriott-under-cma-investigation-for-information-sharing

19 UK CMA Investigates Hotels, CoStar Over Data Sharing - Asian Hospitality, https://www.asianhospitality.com/uk-cma-hotel-data-sharing-probe/

20 UK’s CMA launches investigation into Hilton, IHG and Marriott, https://www.businesstravelnewseurope.com/Accommodation/UK-s-CMA-launches-investigation-into-Hilton-IHG-and-Marriott

21 Europe steps up antitrust enforcement against algorithmic pricing | Loyens & Loeff, https://www.loyensloeff.com/insights/news--events/news/europe-steps-up-antitrust-enforcement-against-algorithmic-pricing/

22 EU rules on concerted practices and agreements between companies - EUR-Lex, https://eur-lex.europa.eu/EN/legal-content/summary/eu-rules-on-concerted-practices-and-agreements-between-companies.html

23 Recent Algorithmic Pricing Developments in the UK and the EU, https://perkinscoie.com/insights/update/recent-algorithmic-pricing-developments-uk-eu

24 The Misleading Consequences of Comparing Algorithmic and Tacit Collusion - European Papers, https://www.europeanpapers.eu/system/files/pdf_version/EP_eJ_2021_2_6_Articles_Luca_Calzolari_00519.pdf

25 Bird & Bird, UOKiK probes software-enabled exchange of strategic information between pharmaceutical wholesalers, 2022 (note: this relates to an earlier proceeding, not the September 2025 algorithmic-pricing investigations), https://www.twobirds.com/en/insights/2022/poland/uokik-probes-software-enabled-exchange-of-strategic-information

26 Algorithmic Pricing Emerges as Enforcement Priority for EU & UK Antitrust Regulators, https://www.morganlewis.com/pubs/2025/10/algorithmic-pricing-emerges-as-enforcement-priority-for-eu-and-uk-antitrust-regulators

27 The Preventing Algorithmic Collusion Act: Strike two - DLA Piper, https://www.dlapiper.com/insights/publications/2025/02/the-preventing-algorithmic-collusion-act-2025

28 New laws regulating algorithmic pricing enacted in New York and California | Davis Polk, https://www.davispolk.com/insights/client-update/new-laws-regulating-algorithmic-pricing-enacted-new-york-and-california

29 On the continuing relevance of state antitrust enforcement in the US - Herbert Smith Freehills, https://www.hsfkramer.com/insights/2026-01/on-the-continuing-relevance-of-state-antitrust-enforcement-in-the-us

30 A Year at the Justice Department’s Antitrust Division | The Regulatory Review, https://www.theregreview.org/2026/03/25/slater-a-year-at-the-antitrust-division/

31 Gig Platforms as Hub-and-Spoke Arrangements and Algorithmic Pricing: A Comparative EU-US Antitrust Analysis - OpenEdition Books, https://books.openedition.org/putc/15512

32 Algorithmic Pricing and AI-Powered Evidence Avoidance: Competition Law Risks and Compliance Strategies | Goodwin - JDSupra, https://www.jdsupra.com/legalnews/algorithmic-pricing-and-ai-powered-1056402/

33 Adapting to and Getting Ahead of Changes in Antitrust and Other Regulatory Demands in 2025 and Beyond - Redgrave LLP, https://www.redgravellp.com/publication/adapting-to-and-getting-ahead-of-changes-in-antitrust-and-other-regulatory-demands-in-2025-and-beyond

34 Synthetic Futures and Competition Law: Towards the Emergence of Precautionary Principle-Minded Approaches - University College London, https://www.ucl.ac.uk/laws/sites/laws/files/cles-6-2024_1.pdf