• by Laurie Sullivan , 5 hours ago

Publishers have been looking for a way to keep their web pages out of AI Overviews, AI Mode, and generative AI features in Google Discover.

Now Google is testing a way that allows publishers to keep their content in traditional AI search results, as measurement companies develop new attribution models for AI-driven search.

A Google Search Console document describes how it began testing this option in June 2026. The control is now rolling out to select website owners.

Help documentation, now available, confirms technical rules that explain how this option works, how long the changes take, and what they do not include.

The decision leaves publishers with a difficult choice. AI Overviews now has more than 2.5 billion monthly active users, while AI Mode has surpassed one billion monthly users, according to Google's estimates.

In June, Google announced it would create an opt-out feature on the same day the U.K.'s Competition and Markets Authority ordered the company to give publishers greater control and transparency over how their content is used in AI search, such as an opt-out option.

Opting out presents both seen and unseen challenges. Publishers lose artificial intelligence (AI) click data, preventing AI Overviews and other AI-related traffic from sending those reading content to their sites.

Google describes the first option announced today as a default, which prevents the content from appearing in Search generative AI features.

Content from other sites will still serve up. The second option prevents a site's content from being visible to users inside those search platforms. Google will not link to the publisher's content or use it to support AI responses.

Controls only affect whether a publisher’s content can appear in certain Search generative AI features, but is not used for ranking or an inclusion signal affecting other parts of Search, according to the documentation.

This will not override a publisher’s other choices to participate in Google’s services such as Merchant Center and Google Ads.

The site's content may still be used to power Google Search -- for example, in helping Google’s systems better understand the language of users’ queries and the language on pages on the web.

Others are working on new attribution models based on whether a publisher chooses to opt-out of AI-related search. 

Matt Gilbert, CEO of Partnerize, which has built a company on AI partnership and affiliate marketing automation, believes the industry needs new attribution models. He calls his company's "HaloIndex." It is part of its VantagePoint attribution tool to track editorial citations in AI search results and connect them to conversions.

Along with the model, Partnerize released a report that identifies 1,056 citation gaps, including 47 high-priority commercial opportunities that represent specific articles and publisher relationships where the gap between influence and compensation is broad enough to act on immediately.

HaloIndex measures the ratio of AI-mediated publisher influence to traditional click-through attribution. It is a comparative measurement designed to quantify the relationship between observed AI-mediated publisher influence and traditional click-based attribution.

A score above 1.0 means clickless publisher influence exceeds what click-based measurement registers. Across the six categories analyzed, the average HaloIndex is 3.84x, driven by the accelerating prevalence of AI-generated answers for commercial search queries.

For publishers, category scores range from 1.54x in Consumer Audio Devices to 10.94x in Luxury Fashion, reflecting the role of research intensity, purchase complexity, and aspirational content in driving AI-mediated discovery.

"Creators are not being compensated as content becomes more important to a brand's ability to be recommended and chosen in a short list in AI search," Gilbert said.

He said publishers and brands are concerned that Google and other companies use content from their sites to train AI models without compensation, even if those systems use their content to make recommendations.

"Visibility alone does not drive recommendations, and recommendations only give a publisher's content the chance to convert," Gilbert said. "The index focuses on verified commercial value and not just visibility. Connecting both is important."