CMA offers guidance on mitigating algorithm risks

The CMA has published a blog post on AI and algorithms, urging businesses to take proactive steps to understand the technology they are deploying and ensure that their use of such tools complies with competition and consumer law.

Much of the blog post repeats what the CMA has said before on this topic - the concerns the CMA outlines in relation to algorithmic pricing are by now well-rehearsed, relating to dynamic pricing and algorithmic collusion (see previous blog posts here and here). The CMA offers advice as to how businesses can mitigate these risks, including ensuring that pricing algorithms neither rely on nor expose competitively sensitive information, exercising caution if using the same tool as a competitor and, where necessary, scrutinising data and algorithms. 

Perhaps more interesting is what the CMA reveals about its approach to reporting and detection and its own growing AI capabilities. In terms of the former, the CMA has recently updated its leniency policy to make leniency available for conduct such as exchanging competitively sensitive information via a shared algorithm. And it offers a reward of up to £250,000 to anyone who tells the CMA about illegal cartel activity, including algorithmic collusion. 

As for its own AI and technical capabilities, since 2019 the CMA has been investing in building its dedicated in-house capability, now brought together in an interdisciplinary directorate, known as “Data, Technology and Insight”. The CMA is developing its screening capabilities to include a range of techniques, including machine learning, AI and agentic systems – and is actively screening for algorithmic collusion and investigating algorithmic systems. It is also continuing to monitor and unpack developments in AI including agentic systems, embedding this insight across its work. 

The CMA is also exploiting advanced technologies beyond digital markets, to detect consumer and competition issues at “unprecedented pace and scale” – including deploying new techniques in data analysis and AI to help detect bid rigging in public procurement, and using agentic AI to identify potential infringements of consumer protection law across the economy, so as to better target enforcement activity where it will have the greatest impact. 

This is an area which is subject to international attention, and authorities are starting to bring enforcement cases – algorithmic collusion is at the centre of the RealPage litigation in the US, and the European Commission has indicated that it has several algorithmic pricing investigations underway. With the CMA making clear it is actively screening for algorithmic collusion, it is a waiting game to see whether and where enforcement action might result.