Back to Blog
Industry Trends June 8, 2026 4 min read

The UK Gives Publishers an AI Overviews Opt-Out. What to Weigh Before Using It

On June 3 the CMA required Google to let publishers keep content out of AI search features without leaving search. What the rule covers, and how to decide whether to opt out.

HR
HBDR Research
June 8, 2026

On June 3, the UK Competition and Markets Authority announced a conduct requirement obliging Google to give publishers effective controls over how their content is used in AI search features. Publishers will be able to keep their content out of features such as AI Overviews, and out of the fine-tuning of Google's AI models, while remaining visible in regular search results. Google must also ensure publisher content is properly attributed, with clear links, when it appears in AI-generated results.

The CMA described the measures as a world first. They build on its earlier designation of Google with strategic market status in general search, which gives the regulator power under the Digital Markets, Competition and Consumers Act 2024 to impose binding conduct rules.

What the requirement says

  • An opt-out from AI features. Publishers get tools to keep content out of AI features in search without leaving search altogether. Until now the practical choice has been all or nothing, because blocking the crawler that feeds search also removes a site from search results.
  • An opt-out from fine-tuning. Publishers can prevent their content being used to fine-tune AI models.
  • Attribution. When publisher content is used in AI-generated results, it must be credited with clear links.
  • Timeline. Google has nine months to implement the changes, though the CMA said it expects important parts of the controls to reach publishers well before that. Google must report on compliance every six months during the first year.

The CMA also indicated it may look further at fair value exchange between Google and content creators. The current requirement gives publishers control; it does not set payment.

Why existing controls fell short

Publishers have had some tools already. Google's Google-Extended robots.txt token lets sites limit use of their content for certain Google AI models, but Google has said it does not govern features in Search. Blocking Googlebot itself removes a site from search entirely, which few publishers can afford. That gap, between an opt-out that does not cover search features and one that costs all search traffic, is what the CMA requirement is meant to close. The details of Google's implementation will determine whether it does.

Why this matters beyond the UK

The requirement covers Google's search services in the UK. But a control built for one major market tends to become a reference point for others, and publishers elsewhere will watch closely how it works in practice. For publishers with meaningful UK audiences, it becomes a real decision over the coming months.

The decision is not obvious

Many publishers' instinct will be to opt out on day one. Before doing that, weigh what you gain against what you give up.

What opting out could gain

  • Leverage. Control over use is the starting point for any negotiation about value. The CMA itself framed the measures as strengthening publishers' negotiating position.
  • Some clicks back. If an AI summary built from your content answers a query without a click, removing your content may return some of those visits to regular results. The effect is uncertain and will vary by query type.
  • Brand control. Your reporting is not paraphrased, possibly imperfectly, inside someone else's interface.

What opting out could cost

  • Visibility in AI features. AI Overviews include links. If your content is cited there today and earns clicks, opting out gives those clicks up.
  • Ceded citations. Competitors who stay in may pick up the citations you leave behind.

How to decide with data

  1. Measure what AI features send you now. Use Search Console and analytics to estimate traffic from queries that trigger AI features and how it has changed over the past year. Segment by content type.
  2. Find content that is being summarized away. Quick-answer and explainer content is the most exposed; original reporting, analysis and tools less so. Your answer may differ by section.
  3. Check how granular the controls are. When Google ships the tools, find out whether they work per site, per section or per page. Granular controls allow testing: opt out a subset and compare it with the rest.
  4. Model the ad revenue at stake. Search visitors are often single-page sessions. Estimate revenue per search session, not just session counts, to see what is at stake either way.
  5. Coordinate with peers. One publisher's opt-out has limited effect on negotiations. Collective positions through trade bodies may carry more weight.

What ad ops should prepare

Whatever editorial and commercial leaders decide, ad operations teams will feel it in the traffic mix. If search referrals shift, pages per session, viewability and revenue per session will move with them. Set up reporting now that isolates search-referred sessions by country and content type, so the effect of any opt-out shows up in weeks rather than months.

It also makes sense to review how search landing pages are monetized. First-screen ad density, lazy loading, and recirculation into a second article matter more when each search visit is harder to earn. A reader who arrives from search and leaves after one page is worth a fraction of one who reads three.

The takeaway

The CMA has given UK publishers something they have long asked for: the ability to say no to AI use of their content without disappearing from search. It is leverage, not a guaranteed traffic win. Measure what AI features send you today, test opt-outs where the controls allow, and treat the decision as a revenue question answered with your own data. Publishers elsewhere should follow the rollout closely, because the UK is now the test case.

Tags: ai search cma google news publishers regulation

Ready to maximize your ad revenue?

Get Started