The UK Competition and Markets Authority (“CMA”) has recently published its strategic update on its approach to artificial intelligence (“AI”). 

This comprehensive update sets out details of:

• the CMA’s understanding of the key risks posed by AI;
• how the CMA intends to address those risks;
• the CMA’s AI capabilities;
• forthcoming changes to the CMA’s enforcement powers;
• how the CMA is working with others on AI issues; and
• the next steps to be taken by the CMA.

While much of the content of this strategic update reflects the concerns set out in the CMA’s recent update paper on AI Foundation Models (“FMs”) which we reviewed in another recent post, it also sets out the CMA’s approach towards the development and deployment of AI more generally. The CMA notes that it plans to continue to research impacts on competition and consumer protection in AI related markets.

Key Risks to Competition 

1. Use of algorithms and similar AI systems

The CMA highlights the potential negative effects of AI systems that affect what choices customers are shown and how they are presented, in particular where algorithms give undue prominence to a particular supplier or platform, rather than the best option for the customer. 

Similarly, the CMA notes that, amongst other things, firms can use AI systems to target customers deemed at risk of switching to their competitors by sending them personalised offers, thereby excluding potential entrants to a sector. This becomes an even greater issue where algorithms target vulnerable consumers or have unfair distributive effects. 

The CMA also notes that firms can use algorithms or AI systems to assist in setting prices, which can facilitate collusion by causing competitors’ prices to rise in tandem with one another, thereby harming the consumer.

2. Foundation Models

FMs are a form of generative AI that are trained on vast amounts of data to produce content for a wide range of tasks and operations. In this latest update, the CMA indicates that its primary concern relating to FMs arise from there being a small number of large, incumbent firms which already control inputs critical to FM development as well as key access points and routes to market for FM deployment. The CMA is concerned that these companies have the power to shape the future development of FM related markets in such a way as to protect their existing market power and extend it into new areas. 

The CMA also identifies a clear need for transparency and accountability regarding the development and deployment of FMs to prevent consumers falling prey to false or misleading information from AI generated content.

CMA’s AI Principles

Following its research into the risks posed by rapid developments in FMs, the CMA published a list of core principles which it set out in its Initial Report of September 2023. These principles were revised in its FM update paper in April this year and are restated for emphasis in its strategic update. 

• Access – ensuring ongoing ready access to inputs to encourage effective challenges from new entrants and ensure existing developers do not become entrenched.
• Diversity – ensuring sustained diversity of business models and model types. 
• Choice – ensuring sufficient choice for businesses and consumers so they can decide how to use FMs.
• Fair Dealing – ensuring there is no anti-competitive conduct by companies, so that the best products and services succeed.
• Transparency – ensuring that consumers and businesses have the right information about the risks and limitations of FMs.
• Accountability – ensuring that FM developers and employers are accountable for FM outputs to maintain the trust and confidence of consumers and businesses.

The CMA urges companies that are active in the development and deployment of FMs to align their business practices with these principles to shape the market in a positive way and lessen the need for future CMA intervention.

How the CMA is enhancing its own AI capabilities

The CMA continues to grow its Data, Technology and Analytics unit, which now consists of more than 80 data scientists, data engineers, technologists, behavioural scientists and digital forensics specialists with a breadth of knowledge about AI technology and the impact it can have on both competition and consumers.

Similarly, the CMA confirms that it has already begun to set up a team of around 70 people within its Digital Markets Unit to exercise its powers under the new Digital Markets, Competition and Consumers legislation (“DMCC”) which will come into force later this year, having received royal assent on 24^th May. 

The CMA also confirms that it has taken steps to incorporate AI into its own systems and practices through its Digital Transformation programme, and has published its own framework and principles for how it should use AI. As part of this programme, the CMA is testing different AI tools and techniques to support its evidence review processes, which form a key part of its case work.

Next steps to be taken by the CMA

The CMA confirms that it intends to take the following actions to address the risks identified above:

• monitoring current and emerging partnerships between companies that control important inputs and/or have strong positions in their respective markets and FMs;
• increasing its scrutiny of arrangements between such companies through the UK merger control regime;
• exercising its powers under the incoming DMCC to prioritise investigations into digital activities that are critical for the development or deployment of FMs, set targeted conduct requirements for firms that have ‘strategic market status’, and impose significant financial penalties for non-compliance; and
• continuing its ongoing investigation into the provision of public cloud infrastructure services in the UK and the impact that future development and deployment of FMs could have on competition in that market.

In addition, the CMA has confirmed that it is working proactively, both internally and with other authorities, to conduct further research into AI, which will inform future guidance and policies. In particular, we may expect to see:

• a paper on AI accelerator chips and their role in the FM value chain;
• joint research with the Digital Regulation Cooperation Forum on consumers’ understanding and use of FM services;
• a joint statement with the Information Commissioner’s Office on the interaction between competition, consumer protection and data protection in FMs;
• a further update on the CMA’s FM-related work in Autumn 2024; and
• guidance for firms on how to comply with consumer law in AI-related markets.

It is clear that there is more to come from the CMA regarding its approach to AI, as it continues to collaborate with other authorities to develop clear policies and guidance on this rapidly evolving sector. With the DMCC coming into force later this year, it is crucial for businesses that use AI in any capacity to pay close attention to these updates.