UK CMA Launches Review of Antitrust Issues in Artificial Intelligence

The CMA's review will focus on understanding foundation models, including large language models and generative AI. Foundation models are trained on vast datasets at scale, often by self- or semi-supervised learning, and can be adapted to a broad variety of end-uses. Those models support a fast-growing range of applications including chatbots, image generation, and code writing. They also will help advance the next generation automation of business processes and workflows.

The stated goal of the CMA's review is to "help this new, rapidly scaling technology develop in ways that ensure open, competitive markets and effective consumer protection." The CMA's announcement follows the UK Government's March 2023 White Paper setting forth the government's high-level policy approach to AI technologies. The UK Government intends to "lead the international conversation on AI" through a "pragmatic, proportionate regulatory approach." The White Paper encourages further engagement with regulators and industry to develop future AI regulations. 

The CMA's work is likely to result in one or more of the following outcomes: (i) recommendations for legislative or regulatory reforms; (ii) a CMA-led market investigation; and/or (iii) guidance to industry about how the CMA thinks antitrust and consumer protection rules should be applied in the AI sector. Antitrust enforcers globally, including in the U.S. and European Union, have spoken about or investigated issues involving AI, but the CMA will be the first major antitrust enforcer to conduct a public investigation focused on application of antitrust rules to AI. As other antitrust enforcers conduct their own studies, they are likely to cite the CMA's work, as they have done for other CMA reports on the tech sector. 

The CMA requested input from industry about how foundation models and their use could evolve, the opportunities and risks of those scenarios for competition and consumer protection, and which principles should guide ongoing development of those technologies. Although some industry participants are likely to receive RFIs from the CMA, key stakeholders should consider whether to submit comments to the extent that regulatory intervention could distort markets or harm their business prospects. 

Companies involved in the following businesses may consider submitting comment: 

• Developers of foundation models;
• Businesses considering developing competing foundation models;
• Downstream users or potential users of foundation models; and 
• Upstream suppliers of infrastructure that underpins AI technologies.

Comments to the CMA are due June 2, 2023. The CMA plans to report on its findings in September 2023.