Google Explores AI Opt-Out for Publishers

Google has confirmed it is developing new controls that would let publishers opt out of having their content used by AI-powered features in Search. The announcement, made on March 18, 2026, is a direct response to pressure from the UK's Competition and Markets Authority — and it signals a meaningful shift in how generative AI interacts with the open web.

This is not a voluntary gesture. It is the product of regulatory force applied through the UK's Digital Markets, Competition and Consumers Act 2024, which gave the CMA new powers to impose conduct requirements on firms with strategic market status. Google was designated with that status in October 2025 — the first company to receive it.

What Google is proposing

In its formal response to the CMA's consultation, Google said it is "exploring updates to our controls to let sites specifically opt out of search generative AI features." Ron Eden, Google's principal for product management, confirmed the direction publicly.

The existing mechanism — the NOSNIPPET meta tag — is a blunt instrument. According to the CMA, using it to block AI summaries also suppresses standard search snippets, reducing organic traffic by approximately 45% based on Google's own research. Publishers have long argued this creates an impossible choice: accept AI summarization of your content, or lose nearly half your search visibility.

Google's proposed update would decouple these controls, giving publishers the ability to opt out of AI features like AI Overviews and AI Mode without sacrificing their position in traditional search results. The specifics — whether this will be implemented via new meta tags, robots.txt directives, or Search Console settings — remain unclear.

The CMA's four conduct requirements

The opt-out proposal is part of a broader regulatory package. On January 28, 2026, the CMA proposed four conduct requirements for Google's search services:

Publisher controls: Content creators must be able to opt out of AI Overviews and AI model training, with proper attribution when their content is used.
Fair ranking: Google must demonstrate transparent, equitable ranking methodologies — including for AI features — with formal dispute resolution processes.
Choice screens: Mandatory search-engine choice screens on Android devices and in Chrome, making it easier for users to switch providers.
Data portability: Enhanced access to Google search data for individuals and businesses.

CMA chief executive Sarah Cardell framed the measures as delivering "more choice and control" for users while providing "a fairer deal for content publishers, particularly news organisations." Google controls over 90% of general search queries in the UK, and more than 200,000 UK firms collectively spent over £10 billion on Google search advertising last year.

What publishers are actually asking for

The publisher response has been pointed. The core demand is granular control — not a single on/off switch, but distinct opt-out mechanisms for different AI use cases.

The Financial Times stated in its response that "publishers must be able to exercise distinct, granular controls over different uses of IP, including grounding, training, and fine-tuning." A publisher might want its content cited in AI Overviews but not used to train Google's foundation models — or vice versa.

The Publishers Association reported a 19% decline in click-through rates to academic reference services, attributing it directly to Google's AI features cannibalizing traffic. The News Media Association pushed for the implementation window to be cut from six months to three.

And there is a non-negotiable line: any opt-out mechanism must not penalize publishers in traditional search rankings. If choosing to withhold content from AI features means dropping in organic results, the opt-out is meaningless. The CMA has explicitly proposed prohibiting ranking penalties for publishers who opt out — but enforcement will be the test.

Google's pushback

Google is not conceding everything. The company warned that some of the CMA's proposed remedies could "have disproportionate consequences" for users or slow innovation. It also argued that full transparency about its ranking systems could "expose our systems to manipulation and abuse," undermining spam-fighting capabilities.

Publishers have called for Google to separate its AI crawler from its standard search crawler, which would give them clearer visibility into how and when their content is being scraped. Google has pushed back on this as expensive and slow to implement — a position the CMA has shown some sympathy toward.

This is the tension at the center of the debate: publishers want control and transparency, Google wants flexibility and speed, and the CMA is trying to find a middle ground that does not break the economics of search for anyone.

Why this matters beyond the UK

The CMA's conduct requirements are specific to the UK, but the precedent is global. If Google builds technical opt-out controls for AI features to satisfy UK regulators, those controls will almost certainly be available worldwide — just as robots.txt became a global standard despite originating from a single proposal.

The EU's AI Act and ongoing antitrust actions against Google in the US create parallel pressure. A working opt-out mechanism in the UK strengthens the hand of regulators everywhere who are watching how AI companies negotiate with the publishers whose content powers their models.

For publishers and content creators, the practical takeaway is straightforward: the tools to control how your content interacts with generative AI in search are coming. The details — how granular, how enforceable, and how soon — are still being negotiated. But the direction is locked in.

Watch for Google's technical implementation proposal in the coming months. The mechanism it builds will shape the publisher-AI relationship for years.

James Calder is the editor of The Search Signal, covering AI-powered search, generative engine optimization, and the future of brand discovery.