CMA Update on Foundation Models and AI Principles: Addressing Risks to Competition and Promoting Accountability and Transparency in the AI Sector

The CMA (Competition and Markets Authority) in the UK has recently published an update paper addressing the potential risks to competition posed by the development of Foundation Models (FMs), which are the technology behind generative AI tools such as ChatGPT. The CMA has identified three key risks to competition in this sector, including concerns about firms restricting access to critical inputs, powerful incumbents distorting choice in FM services, and partnerships reinforcing market power.

In response to these risks, the CMA has set out an updated set of AI Principles that it urges firms to align their business practices with. These principles include ensuring ongoing access to inputs, promoting diversity in business models, providing sufficient choice to businesses and consumers, refraining from anti-competitive conduct, ensuring transparency about the risks and limitations of FMs, and holding FM developers and deployers accountable for their outputs.

The CMA also has several existing workstreams related to AI, including plans to publish a more detailed technical update report, examining critical inputs to FM-related markets, monitoring partnerships closely, and exploring the use of new powers to enforce consumer protection laws in AI-powered markets.

Overall, the CMA is actively monitoring the development of AI technologies and taking steps to ensure that competition in this sector is not stifled. By setting out clear principles and actively investigating potential anticompetitive behavior, the CMA is working to promote fair and competitive practices in the AI industry. Stay tuned for further updates on this important issue.