The Competition and Markets Authority has launched an initial review of artificial intelligence – specifically of foundation models. We summarise the key points here of the CMA's review and how it fits into the UK's White Paper on AI regulation.
What are foundation models?
Foundation models are a type of AI model that is trained on a vast quantity of data, are adaptable for use on a wide range of tasks, and can be used as a base for building more specific AI models. As the UK White Paper on proposed AI regulation (which we wrote about here) recognises, ‘despite high levels of interest in the topic, the research community has not found a consensus on how foundation models work, the risks they pose or even the extent of their capabilities… There is a relatively small number of organisations developing foundation models. Some organisations exercise close control over the development and distribution of their foundation models. Other organisations take an open-source approach to the development and distribution of the technology’.
What will the CMA review?
The initial review will:
- Examine how the competitive markets for foundation models and their use could evolve;
- Explore what opportunities and risks these scenarios could bring for competition and consumer protection;
- Produce guiding principles to support competition and protect consumers as AI foundation models develop.
The review will focus on the questions that CMA considers it is best placed to address – ‘what are the likely implications of the development of AI foundation models for competition and consumer protection?’
How is this connected with the UK White Paper on AI regulation?
This review is part of the CMA’s role to support open, competition markets. It also in line with the UK’s White paper on AI which seeks regulators to consider how their role and regulations support five cross-sector principles: safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.
The CMA expects to gather analysis and then publish a report in September 2023. This is in line with the UK White Paper stating that ‘key regulators’ will be encouraged to publish guidance on how the cross-sectoral principles apply within their remit. It will inform the CMA’s implementation of the Government’s approach to AI regulation, and any recommendations that the CMA may make to the government (including the Office for AI), other regulators (including the DRCF), as well as any guidance to suppliers, developers, businesses and end users.
This should be seen as part of the wider context: AI regulations are coming in multiple jurisdictions – UK, EU, US (federal and state) and others; but they are also diverging, reflecting different regulatory approaches and the contexts & sectors they are trying to regulate. We are actively involved in helping to develop those regulations (for example, we responded to the government's July 2022 policy paper). It is also shortly after the UK government announced £100m investment to ‘kick-start’ a Foundation Model expert taskforce responsible for accelerating the UK’s capability in foundation models.
Sarah Cardell, Chief Executive of the CMA: “Our goal is to help this new, rapidly scaling technology develop in ways that ensure open, competitive markets and effective consumer protection”.