AI and competition law: algorithmic price coordination and market power

Short answer: AI does not change the rules of competition law, but it enables new forms of infringement. If competitors align their prices via algorithms, that can be a prohibited cartel under Article 101 TFEU. And a company with a dominant position on an AI or data market may fall under the prohibition on abuse (Article 102 TFEU). The AI Act governs product safety, not competition — these frameworks are separate.

Algorithmic price coordination

The cartel ban (Article 101 TFEU) prohibits agreements that restrict competition, such as aligning prices. Algorithms can bring this about in several ways:

• Deliberate coordination. Competitors use the same pricing algorithm or pricing software to coordinate prices.
• Hub-and-spoke. A shared service provider or platform synchronises the prices of multiple customers.
• Autonomous alignment. Self-learning algorithms independently arrive at parallel prices — legally the hardest scenario, because an explicit agreement is absent.

The lack of human coordination is no free pass: whoever deploys an algorithm that leads to coordination remains responsible. Competition authorities look at the effect on the market, not only at the programmer's intent.

Pricing software that increases transparency between competitors can also be problematic. Where rivals can track and adjust each other's prices almost in real time through a shared tool, that can effectively switch off price competition.

Market power and data advantage

The prohibition on abuse of a dominant position (Article 102 TFEU) becomes relevant where AI reinforces market power. Access to large datasets, compute and user networks can raise barriers to entry.

Possible risks are self-preferencing, tying of AI services or cutting competitors off from essential data. This connects to the broader debate on market power and concentration in the AI chain.

In addition, the EU has specific platform regulation that applies alongside general competition law and imposes extra obligations on large digital gatekeepers. For AI providers with a strong position it is wise to keep an eye on both tracks.

AI Act and competition: separate tracks

The AI Act sets requirements for the safety, transparency and risk management of AI systems, but contains no competition rules. A system can therefore be fully AI Act compliant and at the same time problematic under competition law — just as compliance is separate from accessibility or other frameworks. Assess both questions separately.

What to do

• Assess your pricing algorithms: can they lead to coordination with competitors, even unintentionally?
• Avoid shared pricing software through which competitors effectively set the same prices.
• Test your market position: do you hold a dominant position on a data or AI market?
• Document independent decision-making on prices and access.
• Involve competition expertise in the design of pricing and platform algorithms.

AI makes coordination easier and faster — and therefore also easier to detect. Testing the design up front prevents efficiency from turning into an infringement.