EU Commission puts generative AI under the competition law microscope

On 19 September 2024, the European Commission (EC) published a Policy Brief on “Competition in Generative AI and Virtual Worlds”. It follows the EC’s January 2024 call for contributions on generative AI and a stakeholder workshop held in the summer.  The Policy Brief doesn’t represent the EC’s official position, but it provides a useful analytical framework that is likely to be reflected in the EC’s future approach to AI-related markets.

Market dynamics and barriers to entry and expansion

The Policy Brief outlines certain “tendencies” of generative AI markets that may be relevant from a competition law perspective, including:

• vertical integration or establishing partnerships to access input resources (such as computing power, data and technical expertise);
• vertical integration or establishing partnerships to access distribution channels;
• the parallel development of open source and proprietary AI foundation models.

The Policy Brief also identifies potential barriers to entry and expansion in relevant markets, focusing on key inputs for development and deployment of generative AI systems such as data, AI accelerator chips, cloud capacity and technical expertise.  Other potential barriers – which in the view of the authors are “quite different” from the characteristics of digital markets more generally – include the high initial fixed costs of pre-training foundation models and the high marginal costs of operating such models.

Theories of harm

The Policy Brief goes on to discuss several theories of harm that may be relevant in generative AI markets (i.e. reasons why particular types of conduct may give rise to competition law concerns). These include:

1. Agreements between competitors that reduce competitive constraints or facilitate exchange of competitively sensitive information;
2. Vertically integrated firms adopting pricing policies aimed at “margin squeezing” competitors;
3. Large incumbent firms granting third parties exclusive access to key inputs (e.g. computing infrastructure and high-quality datasets) or preventing competitors from accessing them;
4. Large incumbent firms using their market power to limit choice or distort competition in downstream markets when distributing and commercialising AI applications.

In connection with point 4, Competition Commissioner Margrethe Vestager emphasised in a speech on 28 June that “[…] big tech companies could make it difficult for smaller foundation model developers to reach end users. Whether alone or in alliances with preferred partners. So we are closely monitoring distribution channels to make sure businesses and consumers still have a wide range of choices among foundation models”.  The Policy Brief notes that the EC is currently investigating whether agreements between Google and mobile device manufacturers (such as Samsung) for the pre-installation of Google’s small AI model “Gemini Nano” on mobile devices may raise competition concerns by making it more difficult for other foundation models to be accessed on those devices.

To address the competition risks, the Policy Brief advocates vigorous antitrust, merger control and Digital Markets Act (DMA) enforcement. This reflects comments made by Commissioner Vestager in her 28 June speech:

“[…] Now is the time to act. Strong competition enforcement is always needed at times of big industrial and tech changes. It is then that markets can tip, that monopolies can be formed, and that innovation can be snuffed out. When competition enforcement steps in, it allows different business models and new ideas to develop. […] AI and the metaverse are developing at breakneck speed. We cannot just sit back and see how things pan out.”

Multi-jurisdictional scrutiny

The European Commission is not the only competition authority for whom generative AI is a current policy and enforcement priority.  In July 2024 the US Department of Justice (DoJ), the US Federal Trade Commission (FTC), the UK Competition & Markets Authority (CMA) and the EC issued a joint statement on competition in generative AI foundation models and AI products, signalling the agencies’ joint commitment to preventing anticompetitive practices in this area.

It has also been reported that the FTC and DoJ opened antitrust investigations into Microsoft, OpenAI and Nvidia in June 2024, focusing on the companies’ business practices in AI-related markets. In July, Reuters reported that the French Competition Authority had also launched an investigation into Nvidia over alleged anticompetitive practices. Nvidia is a leading supplier of chips that train and run AI systems.

In our next post, we will take a look at recent CMA activity in this area.