CMA imposes 'world-first' AI search controls on Google in the UK

The Competition and Markets Authority (CMA) has imposed a new conduct requirement (CR) on Google search. The CR follows the CMA’s 2025 decision to designate Google as having strategic market status in general search services, enabling the regulator to impose targeted rules designed to promote fair dealing, open choices, and trust and transparency.

The headline change is a 'world-first' requirement giving publishers the option to stop their content being used to power Google’s AI search features, including AI Overviews and AI Mode.

Google must provide publishers with:

• 'Effective controls' to decide how far their content is used across the full range of AI-related search applications (including whether content used for general search can also be used for AI training, such as fine-tuning and grounding). The CMA expects these controls to be implemented at both directory and page level.
• Detailed metrics on user engagement with their content in search generative AI features. This includes user impressions (where the content is attributed in a search generative AI feature), user engagement (including data on click-throughs to the publisher's website) and click-through rate. The data must be provided for each publisher and disaggregated from other elements of general search.
• Effective attribution, with end users provided with links to publisher content in AI-generated search results.

Google has nine months to implement the full set of changes, although it has recently announced a rollout of new features for testing. The CMA has invited feedback from interested parties as the changes are rolled out.

The CR benefits any publisher whose online content is made available to end users in the UK using Google's general search services.

What the new Google AI Overview controls mean for publishers

Publishers have generally accepted Google’s access to their content for the purposes of display and linking within search results. The use of that content in generative AI features, however, raises distinct issues and reduced website traffic.

Publishers have also raised concerns in the context of the UK Government’s wider policy work on AI and copyright. The CMA has nevertheless been clear that the CR is confined to addressing issues arising from Google’s strategic market status in general search services and operates independently of that broader policy work on copyright and creator control.

One objective of the CR is to strengthen publishers’ bargaining position when negotiating content arrangements with Google, particularly in the case of news organisations. Larger publishers have, in some cases, been able to agree terms with leading AI developers for the use of their content in AI training. By contrast, smaller publishers often lack visibility into how their content has been used. The CR is intended to improve transparency and afford greater control, at least in relation to the use of content within Google’s AI Overviews.

That said, increased user reliance on AI Overviews within Google Search may, in practice, limit publishers’ willingness to opt out of such use. There is therefore a concern that the CR may not, on its own, sufficiently strengthen publishers’ bargaining position. The CMA has indicated that it may introduce further measures, if necessary, to ensure a fair exchange of value between Google and publishers.

This publication is a general summary of the law. It should not replace legal advice tailored to your specific circumstances.

© Farrer & Co LLP, June 2026