The UK’s Competition and Markets Authority (CMA) launched a review on May 4 to assess the potential competitive implications of rapidly emerging artificial intelligence (AI) foundation models—a type of AI technology trained on vast amounts of data that can be adapted to a wide range of tasks and operations (e.g., AI-powered chatbots).
As AI transforms various sectors globally, governments and regulators have been reacting to the risks and opportunities it presents. The CMA’s review comes in light of recent investigations into data privacy issues related to AI chatbots by data protection authorities in Italy, Canada, and South Korea, resulting in a temporary ban in Italy (reversed only a few days ago).
Earlier this year, and in response to these developments, the UK Department for Science, Innovation and Technology (DSIT) published a white paper proposing a context-specific approach to AI regulation. The white paper outlined five guiding principles for AI used in the United Kingdom and called for a flexible approach, allowing existing expert regulators such as the CMA to use the tools currently available to them and issue practical guidance to organizations.
As part of its review, the CMA will consider the recently published white paper and its proposed AI framework, for which the UK government seeks feedback by June 21, 2023.
SCOPE OF THE CMA’S REVIEW
The CMA’s review seeks to map the effects that AI foundation models are likely to have on the structure and functioning of competition (both in AI-related and other markets), as well as the potential antitrust and consumer protection risks posed by them. The CMA’s review is not targeting specific companies but will engage with a variety of stakeholders, including businesses, academics, and other interested parties, to gather comprehensive information.
The CMA’s review is expected to focus on the following issues:
- Market Concentration: The extent to which markets for AI foundation models are and can be sufficiently competitive, in light of the technology, amount of data, and economies of scale required for their successful development and deployment, all of which might constitute significant barriers to entry precluding the emergence of competition
- Vertical Integration: The extent to which the expanding influence of AI in multiple sectors could result in vertical integration solidifying the market power of dominant firms
- Risk of Exclusionary Conduct: Whether dominant firms controlling AI foundation models will have the ability and incentive to engage in exclusionary conduct, such as limiting access to essential AI technology or data, stifling competition and innovation
- Risk of Collusion: Whether AI foundation models may facilitate collusion between competitors by automating pricing decisions, potentially resulting in higher prices and reduced consumer choice
POTENTIAL OUTCOMES & IMPLICATIONS
The CMA stated that it will publish a short report outlining its findings in early September 2023. The swift launch of a review, largely in response to the rapid evolution and widespread use of powerful AI chatbots, signals the emergence of AI-specific antitrust regulation and enforcement, potentially involving
- specialist CMA guidance concerning competition law compliance by firms developing, deploying, or using/subscribing to AI foundation models;
- increased regulatory oversight of AI development, deployment, and market practices through the CMA’s existing powers to conduct market studies, and investigate suspected anticompetitive agreements and abuses of market dominance;
- heightened CMA scrutiny of M&A transactions involving firms developing or deploying AI models, in particular those that are likely to result in vertical integration;
- a possible designation of firms active in AI development and deployment as having “strategic market status” (SMS) in respect of a digital activity linked to the United Kingdom. Under the United Kingdom’s draft Digital Markets, Competition & Consumers Bill, an SMS designation would subject such firms to tailored Codes of Conduct, as well as the possibility of “pro-competition interventions” by the CMA’s Digital Market Unit; and
- The CMA’s review, alongside similar efforts by the US Federal Trade Commission and other regulators, may lead to greater international cooperation in regulating the AI industry and in enforcing competition law in respect of suspected violations with cross-border elements.