Algorithmic Collusion In Competition Law

Author: Shreya Lode, Rashtrasanta Tukdoji Maharaj Nagpur University.

To the Point

The rise of artificial intelligence and algorithm-driven marketplaces has transformed the competitive landscape in India. While pricing algorithms help firms optimize business strategies, they also create the possibility of algorithmic collusion — coordinated pricing behavior without direct human agreement. This new form of collusion challenges traditional competition law, which was designed to detect human intention and communication, not machine-generated parallelism.. With the Competition Commission of India (CCI) increasingly engaging with digital markets, the need to examine whether the Competition Act, 2002 is adequately equipped to tackle algorithm-facilitated anti-competitive conduct is more pressing than ever. This article explores how algorithmic collusion arises, why it fits awkwardly within Indian legal doctrine, and what regulatory reforms may be necessary for technology-driven markets.

Abstract

This paper critically examines the emergence of algorithmic collusion in India through the lens of competition law. It addresses the legal ambiguity surrounding whether coordinated pricing achieved through algorithms amounts to a cartel under Section 3 of the Competition Act, 2002. Using doctrinal analysis of statutory provisions, comparative jurisprudence, and academic commentary, the discussion argues that current regulatory tools struggle to establish intent, liability, and enforcement mechanisms. The paper concludes by proposing legislative and technological reforms to strengthen India’s competition framework in the age of artificial intelligence.

Use of legal jargon

Algorithmic collusion refers to concerted market behavior facilitated through artificial intelligence or pricing algorithms, rather than explicit communication among firms. Under Section 3(3) of the Competition Act, 2002, cartels involving price-fixing, bid-rigging, or market allocation are presumed to have an appreciable adverse effect on competition (AAEC).
Key issue for happening of algorithmic collusion are:
Tacit collusion
It is a form of anti-competitive behavior where competitors in a market secretly coordinate their actions to raise prices or restrict output without an explicit agreement
hub and spoke cartel
This is an anti-competitive arrangement where competitors (“spokes”) coordinate their actions indirectly through a central intermediary (“hub”). Unlike traditional cartels with direct communication, the hub facilitates collusion, such as price-fixing or market allocation, by enabling an indirect “agreement” among the competing spokes, often seen in modern contexts like online marketplaces.
Cartelization:
Cartelization refers to a situation where competing firms or market participants collaborate instead of competing, with the intention of controlling prices, output, supply, or market share. It is an anti-competitive agreement and is illegal in most jurisdictions

The Proof:

The legal foundation of cartel enforcement in India rests primarily on Section 3 of the Competition Act, which prohibits anti-competitive agreements. Section 3(1) is broad enough to cover tacit collusion, but Section 3(3) focuses on intent-based coordination. When an algorithm independently adjusts prices in response to competitor signals, the line between conscious collusion and automated equilibrium blurs. The CCI must prove either an agreement or a concerted practice, but algorithmic conduct may lack documentary trails, communication, or human instruction.
Three major gaps arise:
Absence of explicit agreement – Algorithms may react to shared data or common software, producing collusive outcomes with no communication.

Difficulty in attributing liability – Should accountability lie with the developer, the firm using the algorithm, or the algorithm itself?

Detection challenges – Parallel pricing patterns may reflect competitive optimization rather than conspiracy.

Thus, while the Act technically covers collusion, enforcement becomes practically impossible without new interpretative tools or statutory amendments.

Case Law
Samir Agarwal Vs. Competition Commision Of India & others.
Civil Appeal No. 3100 of 2020:
In Samir Agrawal v. CCI & Others (2020), the informant alleged that Ola and Uber used algorithmic pricing that amounted to a hub-and-spoke cartel and resale price maintenance under Section 3 of the Competition Act. The key legal issues were, whether an individual with no direct commercial interest had locus standi to file information, and whether algorithm-driven pricing constituted anti-competitive agreement or collusion. The facts centered on drivers not setting prices themselves, with the platform’s algorithm determining fares. The Supreme Court held that any person can provide information to the CCI (wide locus standi), but upheld the CCI and NCLAT’s decision that there was no prima facie evidence of collusion among drivers or a hub-and-spoke cartel, and that algorithmic pricing alone does not amount to price-fixing without proof of an agreement. The Court concluded that the case should be dismissed on merits, though locus standi is broad. After the decision, the major impact was clarification, not a statutory change, that digital-market complaints can be filed by any person, and competition authorities must evaluate algorithmic pricing cautiously, requiring clear evidence of concerted action before treating platform algorithms as cartel-facilitating mechanisms. NCLAT in its order, focused on the absence of connectivity between drivers inter se.

Suo Motu Case No. 03 of 2015
In Re: Alleged Cartelization in the Airlines Industry
This emanated upon the receipt of Letter in Lok sabha. CCI told DG to go through an investigation. It was observed that despite differences in base fares and airlines fuel surcharge, the end fares charged by all the airlines for tickets were almost similar. Later, they viz DG was told for supplementary investigation and came out:
There was no evidence of cartelization or collusion among the airlines.
While “price parallelism” (i.e. similar fare levels across airlines) was observed, the CCI held that such parallelism — by itself — does not prove an agreement or collusion. The agencies’ price-setting decisions appeared to be independent.
The pricing algorithms used by different airlines were not common or shared; each airline used its own software, with inputs that reflected its own historical data, capacity, flight frequency, demand, and other route-specific variables.
The final decisions on ticket fares were made by each airline’s own “revenue-management teams” (human analysts), not automatically imposed by a common algorithm. The software’s role was limited to assisting these teams in revenue optimization.
Also, there was no evidence of communication or information exchange between airlines that would indicate collusion. On 22 February 2021, CCI officially closed the investigation with a clean-chit to the airlines — finding no contravention of Section 3 (antitrust/cartel provisions) of the Competition Act.

Conclusion

As India rapidly integrates artificial intelligence into its social and economic systems, it becomes essential to ensure that these technologies support a fair, stable, and prosperous society. The current Competition Act, 2002 — particularly Sections 3 and 4 — was not designed to address AI-driven market behaviour. Algorithmic collusion, in particular, poses a silent and scalable threat, operating as a new form of cartelization that often escapes traditional enforcement methods. While the Act broadly prohibits price-fixing, its assumptions of human intention and communication are outdated in a digital marketplace.
To safeguard market integrity, India must introduce proactive reforms: clearly defining algorithmic liability under Section 3, establishing specialised AI-monitoring units within the CCI, imposing mandatory transparency rules for pricing algorithms, and creating regulatory sandboxes to audit algorithmic behaviour. Without such evolution, the risk is that algorithms — rather than businesses — will begin to control market prices, eroding consumer welfare and weakening the core principles of competitive markets.

FAQS:

Is Algorithmic pricing not equal to Algorithmic collusion?
No, Algorithmic Pricing is not equal to Algorithmic Collusion
Algorithmic Pricing
It means companies use software or AI to set prices dynamically based on demand, supply, consumer behaviour, and historical data.
Example: Surge pricing by cab platforms, discounts adjusted automatically on e-commerce sites.
Algorithmic Collusion
This happens when algorithms, intentionally or unintentionally, start aligning prices with competitors — reducing competition, raising prices, and harming consumers. Here, AI ends up functioning like a digital cartel

What reforms are needed?
India must strengthen its competition law to deal with AI-driven market behaviour. First, the CCI needs a specialised investigation unit with power to analyse algorithms, access internal digital systems, and audit pricing software. Second, the Competition Act, 2002 must be amended to include a clear definition of algorithmic collusion, recognising machine-facilitated coordination even without human intent.
Mandatory transparency rules should require firms to disclose algorithm architecture, data inputs, and third-party pricing tools. Regular AI-audits must monitor market behaviour and flag suspicious pricing patterns. Finally, liability should shift from proving intent to evaluating outcome, ensuring companies remain responsible for their algorithms even when collusion is unintentional.

Is CCI really taking steps in order to stop this collusion?
Yes, even though it is, they are saying that companies that can collect, analyse and leverage vast amounts of data can gain a major competitive advantage. They are taking necessary steps towards this issue such as Settlements and commitment, Deal value threshold for new mergers and acquisitions, Leniency plus regimes and introduction of Hub and spoke model in the statute book, also they are conducting market study on AI and Self-Audit Framework.

References