Key Points
- Since generative AI went mainstream in 2023, competition regulators on alert for have potential competition issues in the sector have been preparing for possible enforcement actions.
- Regulators around the globe are examining the competitive landscape of the AI supply chain as well as partnerships between startups and established businesses that operate at different levels of this supply chain.
- Regulators have been clear that they will not hesitate to use enforcement powers to keep emerging markets contestable.
- Issues such as access to key inputs and vertical integration are being closely monitored, yet much remains uncertain about how the AI supply chain will develop and which competition issues will gain the most attention.
Artificial intelligence (AI) — and in particular generative AI, which creates new content — has recently emerged as a transformative tool with the power and potential to reshape familiar industries. Although disruptive technologies can bring benefits and be pro-competitive in many instances, regulators around the globe are watching over this rapidly developing space to ensure that it evolves in a way that benefits competition and consumers.
Increasing Investigative Reviews
Drawing on experience gained in the digital sector, regulators have been quick to proactively explore and prepare for the challenges stemming from AI as the sector takes shape.
In the U.S., the Federal Trade Commission (FTC) highlighted the competition issues associated with generative AI in a June 2023 blog post. Initial reviews that identify potential risks to competition in generative AI have already been carried out by the:
- U.K.’s Competition and Markets Authority (CMA) (September 2023).
- Australian Competition and Consumer Commission (October 2023).
- Portuguese Competition authority (November 2023).
Reviews are currently underway by the European Commission (EC) and Hungarian and French competition regulators.
Partnerships Under Scrutiny
Regulators are also scrutinizing agreements between large digital players and nascent AI startups. A number of regulators are investigating the impact of these partnerships on competition as part of the broader reviews of AI outlined above. In January 2024, the FTC launched a formal inquiry into some of the main partnerships, ordering the parties to provide information.
Due to their structure, these relationships do not always fall clearly within the scope of merger control rules. Therefore, a number of competition regulators are considering whether they have jurisdiction to review Microsoft’s investment in OpenAI, for example.
- Germany’s competition regulator concluded in November 2023 that the cooperation did not fall within the purview of the German merger control regime but did not rule out that it could become notifiable if Microsoft’s influence over OpenAI increased.
- Launched in January 2024, the EC’s evaluation of whether the cooperation might be reviewable under European merger control rules is ongoing.
- The CMA, which has a more flexible jurisdictional test than the EC, is considering whether to launch a formal merger investigation following its December 2023 invitation to comment on the partnership.
Areas of Potential Concern
While acknowledging that the technology is still fast-evolving, comments from regulators point to a number of areas of potential concern across the AI supply chain.
Access to essential inputs. The main competition focus is directed at the AI building blocks: data, computational resources (specialized chips and cloud computing) and labor/talent. Regulators are closely monitoring incumbent businesses that control any one of these essential inputs to ensure the promotion of a level playing field for new entrants. This includes cloud providers, companies with extensive data sets accumulated through providing services in adjacent markets, companies with access to unique and essential data sets (such as health care or finance), and companies with incentives to lock in skilled workers.
- Regulators are paying particularly close attention to large digital players at different levels of the AI supply chain, with some questioning whether such vertical integration could give rise to issues around discriminatory access to key inputs (including pricing and terms of use) and exclusive dealing, or whether companies might use generative AI applications to favor products or services within their ecosystem.
- A related concern is that companies will use mergers and acquisitions or other arrangements to entrench market power. The cloud service sector, which has faced significant antitrust scrutiny in the last two years, has been identified as a key area to watch in this regard. Many cloud service providers are not only developing proprietary AI models but also making large investments in leading AI developers (such as Microsoft’s investment in OpenAI). While such partnerships enable new entrants to scale up and compete, regulators are concerned that a handful of companies dominate the cloud services market and that, as AI continues to advance, partnerships between cloud service providers and AI startups may intensify the trend toward concentration and incentivize these providers to exclude downstream competitors.
- A further concern is that companies operating in adjacent digital platform services markets might employ exclusionary strategies including linking new generative AI products with existing core products such as search engines, operating systems or productivity software (tying), or offering bundled packages of multiple products (bundling), potentially distorting competition.
Open-source models. While open-source generative AI models promote innovation and new entry, some regulators are concerned that this format will not encourage the development of models that can rival more closed models. The risk of companies using open-source models to draw in users and then locking in customers and locking out competition through commercial and technological practices has also been identified as a potential problem.
Looking Ahead
Regulators seek to shape these emerging markets and have been clear that they will intervene with formal enforcement powers where appropriate.
Merger control may be used as a tool to address concentration risk at different levels of the AI “stack,” or supply chain. Transactions that are likely to come under scrutiny include:
- Horizontal transactions by which incumbent digital platforms seek to acquire emerging AI competitors.
- Vertical transactions involving a player that controls access to a key AI input such as data that is not widely available.
There is also likely to be continued discussion about whether existing merger control regimes can be used to review partnerships between incumbent digital players and AI businesses.
New regulatory regimes may also be deployed. While the EU’s Digital Markets Act was not initially intended to apply to generative AI, there have been several calls for the rules to be updated in order to explicitly encompass this growing technology. In the U.K., the CMA is expected to receive new powers later this year to regulate the digital sector under a regime that would be sufficiently flexible to examine AI-driven markets.
Market studies are another tool that some regulators may use to tackle anticompetitive features at a marketwide level. Notably, the CMA’s ongoing market investigation into cloud services may consider the AI supply chain from a competition perspective.
Finally, traditional antitrust rules may be used to tackle issues such as bundling and self-preferencing.
Professional support lawyer Elizabeth Malik contributed to this article.
This memorandum is provided by Skadden, Arps, Slate, Meagher & Flom LLP and its affiliates for educational and informational purposes only and is not intended and should not be construed as legal advice. This memorandum is considered advertising under applicable state laws.