CMA Requires Google to Let Publishers Opt Out

The UK Competition and Markets Authority (CMA) imposed a new conduct requirement on Google for its general search services, published on June 3, 2026, requiring Google to give publishers tools to prevent their content from being used in AI features such as AI Overviews and AI Mode. The CMA press release states the requirement also lets publishers opt out of their content being used for the 'fine-tuning' of AI models and requires clearer attribution with direct links when publisher content appears in AI-generated search results (gov.uk). Reuters and the BBC report the CMA had designated Google with strategic market status, the step that enables these targeted conduct requirements because of Google's dominant share of UK search traffic, which the regulator and reporting put at more than 90% (Reuters; BBC; gov.uk). Google said it is already 'exploring updates to our controls to let sites specifically opt out of search generative AI features' (Press Gazette).

Editorial analysis - technical context: Public reporting describes the opt-out as applying to generative features that produce summaries and conversational responses, not to the traditional organic ranking algorithm. Reuters and the BBC note Google told publishers the controls would prevent sites that opt out from receiving impressions or traffic from generative AI results while leaving standard search rankings unchanged. The CMA press release also requires Google to improve how it attributes publisher content inside AI responses, for example by providing clearer links back to original articles (gov.uk; Reuters; BBC).

News publishers and trade groups have argued that AI summaries reduce click-through rates by answering queries without a user visiting the original article, a dynamic documented in coverage by Reuters and the BBC. The CMA frames its action as strengthening publishers' bargaining power and protecting the economic value of journalism; the BBC and gov.uk quote CMA Chief Executive Sarah Cardell calling the requirement a 'world-first' and saying it will provide 'fair treatment, greater transparency and meaningful choice for businesses and consumers' (gov.uk; BBC). Trade bodies welcomed the intervention: Dan Conway, chief executive of the Publishers Association, called the requirement, which bars Google from using publishers' content to train its AI models without consent, 'a significant step in the right direction,' while the News Media Association, which represents UK national and regional publishers, had pressed for and welcomed the controls (Publishers Association; Press Gazette).

Editorial analysis - practitioner implications: The opt-out for fine-tuning explicitly creates a regulatory mechanism separating content licensing for surface-level summarization from downstream model training. For teams tracking data provenance and training-set composition, this increases the likelihood that content providers will demand explicit agreements or technical controls that can exclude material from training pipelines. Observers should expect increased demand for publisher-level controls and clearer signals (for example, robots.txt extensions or new headers) to indicate permitted uses, although the CMA text does not prescribe a specific technical mechanism (gov.uk; Search Engine Journal; 9to5Google).

For practitioners, monitor three indicators over the coming months:

• •the exact technical implementation Google deploys for opt-outs and whether it uses metadata, webmaster-tools toggles, or contractual and licensing flows (reported testing by Google is noted in Reuters and 9to5Google)
• •publisher adoption rates and any coordinated industry response or licensing negotiations
• •measurable effects on traffic and referral patterns from AI Overviews compared with traditional search links, which Google says it will provide new insights on to publishers (Reuters; 9to5Google)

The CMA gave Google nine months to implement the changes, said it expects key controls to reach publishers before that deadline, and will require regular compliance reports, with further action possible if needed (gov.uk; Press Gazette).

What is reported: the conduct requirement and its headline provisions are documented in the CMA press release and covered by Reuters, BBC, WSJ, and other outlets. What is not reported: detailed technical designs for opt-out controls, the timetable for any global rollout, and any commercial terms Google may offer publishers; public coverage indicates Google is testing features in the UK first but does not provide final product specifics or contractual frameworks (gov.uk; Reuters; BBC).

Regulators in major markets are now explicitly intervening in how search engines surface third-party content inside generative AI features, and the Guardian notes the UK move could have global ramifications. For practitioners, the ruling raises the importance of robust tracking of referral traffic, clearer content-use metadata, and readiness to respond to publisher control signals in both indexing and any downstream training pipelines.