Predictable Pricing, Unpredictable Liability- AI Algorithms and Tacit Collusion Ahead of the Indian Competition Law: AUTHOR: Navya Chauhan
Pricing algorithms can produce stable, competitive prices without any interaction or communication between rivals in a market. The Competition Act, 2002 is structured upon an “agreement” that leaves a lacuna for algorithmic coordination. This article examines the statutory framework and the Hon’ble Supreme Court’s treatment of such AI-based algorithms. It argues that India needs a comprehensive framework to address shared pricing tools, diligence for users and sellers, including an improved technical enforcement mechanism.
ARTICLE
Abstract
Pricing algorithms can produce stable, competitive prices without any interaction or communication between rivals in a market. The Competition Act, 2002 is structured upon an “agreement” that leaves a lacuna for algorithmic coordination. This article examines the statutory framework and the Hon’ble Supreme Court’s treatment of such AI-based algorithms. It argues that India needs a comprehensive framework to address shared pricing tools, diligence for users and sellers, including an improved technical enforcement mechanism.
Keywords:
Algorithmic collusion, tacit collusion, cartels, Competition Act, Samir Agrawal case.
Introduction
It is pertinent to firstly note that parallel pricing is not unlawful in India. The Competition Act, 2002 penalises collusion, but not interdependence. But AI-based models that re-price in seconds, monitors rivals constantly and rigorously learns from outcomes and market trends can sustain collusive prices without any check or liability. The Indian legal system currently has no legal framework to monitor or regulate the same, let alone penalise it.
Statutory Framework
Section 3(1) of the Competition Act, 2002 prohibits agreements causing or likely to cause an appreciable adverse effect on competition (AAEC). Section 3(3) further presumes appreciable adverse effect on competition for horizontal agreements that fix prices, limit supply, allocate markets or rig bids. Additionally, Section 2(b) of the Act defines “agreement” broadly to include any arrangement, understanding or action in concert, whether or not formal. The contravention for the same is punishable further under Sections 27 and Section 26. The roadblock is thereby apparent, wherein parallel conduct has needed certain factors such as communication or information exchange before it can be inferred. A self-sufficient, independently functioning AI-algorithm, however, may leave no such trace.
Indian Jurisprudence
In the case of Samir Agrawal v. ANI Technologies (CCI Case No. 37 of 2018), the informant alleged that Ola and Uber were hubs and their drivers fell under the ambit of Section 3(3)(a) as a price fixing cartel. The Competition Commission of India (CCI), NCLAT and the Hon’ble apex court, however, found no such contravention. The CCI gave its reasoning by stating that a hub and spoke arrangement involves competitions exchanging sensitive information about a third party, which was not the case here. The court set aside NCLAT’s finding that the informant lacked the locus standi and stated that Ola and Uber used different, independently developed algorithms, which distinguish the case from rivals adopting a single shared pricing tool. However, the Hon’ble court did not address anything pertaining to common vendors or autonomous convergence.
The 2023 Amendment
The Competition (Amendment) Act, 2023 amended Section 3(3), stating that an enterprise not engaged in identical or similar trade is presumed to be a part of a cartel if it participates or intends to participate in furthering it. This widened the ambit and scope of the provision to include software vendors coordinating rivals, but it still presupposes the presence of an agreement among the two parties, and requires proof of participating or intent, which is rendered difficult and untraceable in the case of an AI-based algorithm model.
CCI’s Market Study on AI
The Competition Commission of India published a market study on AI and its effect on competition on 6 October 2025, which warned that AI pricing algorithm systems can create new kinds of collusions in the market without human intent or communication, and that such algorithms can blur the line between explicit and tacit collusion. Around 37% of the surveyed startups were concerned about AI-facilitated collusion.
The CCI, in its report, also gave recommendations including documenting AI-based decision-making, periodically reviewing algorithmic outputs to eliminate inadvertent collusion and reviewing AI-driven pricing strategies.
However, this report is not binding and its recommended remedy was essentially just self-audit, creating no legal framework for regulation and surveillance. It was thereby, criticised for such a light response. Compliance recommendation cannot create any liability upon the perpetrators that the Act does not already impose.
The Question of Liability
Even if coordination is established, attribution is still a grey area in Competition Law. Section 27 of the Act pertains to penalties, which can reach three times the profit or 10% of turnover for cartels. However, where one competitor’s model sets prices for several competitors, each user may claim it merely adopted an industry tool. The 2023 amendment to Section 3(3) helps against facilitators but it depends on proof of intent and participation. Section 48 extends liability to persons in charge of a company, which raises a further question: whether a manager can be said to have “intended” an outcome that an independent AI reached on its own? The Courts still need to formulate and give directions for this question.
Way Forward
The present legal framework lacks the inclusion of AI-based pricing tools in its ambit. The CCI should state that the shared competitor tools fed with rivals’ non-public data raise an interference of concert. Further, users and competitors should be able to demonstrate what data a tool uses and how it produces an output. The CCI should strengthen its technical capacity in order to audit algorithms and regulate supervision and surveillance since tacit collusion is unlikely to overtly surface.
Conclusion
The increasing use of AI-based pricing algorithms in the market by almost every competitor makes it imperative for the Indian legal system to bridge the gap left by the Samir Agrawal case, which rejected a hub and spoke theory on facts, and the 2023 amendment, which targets facilitators within agreements, along with CCI’s study on self-audit. The contemporary situation demands that the CCI defines which algorithmic designs and date practices attract liability so that predictable pricing does not carry unpredictable exposure.
References
· SCC Online
· HeinOnline
· The Competition Act, 2002
· The Competition (Amendment) Act, 2023
· https://www.cci.gov.in/antitrust/orders/details/228/0
· Algorithmic Pricing and Cartelisation: Is the Competition Act Equipped for AI-Driven Collusion?, Primelegal, 22 August 2026
