Direct answer: Publishers can now decide through Google Search Console whether their sites appear in and help ground Google’s generative Search features. The professional response is not a reflexive opt-in or opt-out. It is a documented participation decision backed by baseline data, clear ownership, a test period, and a scheduled review.

For years, publishers treated Google visibility as a mostly technical question: can the page be crawled, indexed, understood, and ranked? Generative search adds a new layer. A publisher can now make a separate governance decision about whether its site should appear in AI Overviews, AI Mode, and AI Overviews in Discover.

Google announced the control on June 3, 2026 and updated the announcement on August 31 to say the features had rolled out to websites worldwide. Google also says opting out of the covered generative features will not be used as a ranking signal for ordinary search results outside those features.

That changes the question from “How do I optimize for AI search?” to something more fundamental: What is our policy for participating in AI search, and how will we know whether that policy is helping the publication?

What changed for publishers

The new Search Console control is designed to let website owners decide whether their sites can appear in and help ground responses in Google’s generative Search features. Google is also surfacing new performance insights, including impressions, pages that appear in generative responses, and country-level information.

That matters because publishers can finally connect a policy decision to measurable evidence. A news site may value reach and branded discovery. A subscription publication may care more about registrations. A professional publisher may care about newsletter signups, book sales, leads, or licensing inquiries. A reference site may prioritize authority and citation visibility.

The correct setting can therefore differ by publication. What should not differ is the need to document the decision.

The Search Console switch is not a copyright license

Search visibility, model training, copyright permission, indexing, and contractual licensing are related but distinct issues. A Search Console participation setting does not determine who owns the work, whether a model may train on it, or whether a third party has a contractual right to reproduce it.

Google itself distinguishes the generative Search control from tools such as snippet controls and Google-Extended. Publishers should do the same internally. Discovery settings belong in a distribution record. Ownership and permissions belong in a rights record.

That distinction connects directly to the rights-provenance standard. If a publisher cannot identify the owner, registration status, licensed uses, and source history of a work, changing a search toggle does not solve the underlying rights problem.

Before changing the setting, capture a baseline

A publisher should never make a visibility decision without knowing what “before” looks like. Capture at least 28 to 30 days of baseline data when possible. Record ordinary organic search visits, branded search activity, newsletter subscriptions, lead submissions, book or product conversions, page-level revenue where available, and the pages that already attract meaningful traffic.

Then export or screenshot the generative Search information available in Search Console. The goal is not to chase a single metric. The goal is to understand whether generative visibility creates useful downstream behavior.

The 30-day publisher test

Use a fixed test window instead of reacting to a few volatile days. Record the effective date, the person authorized to change the setting, the business goal, the success threshold, and the rollback condition. Review once around day 14 for obvious technical problems, then again at day 30 for an actual decision.

For larger publishers, segment the analysis by content class even when the platform setting applies broadly. Evergreen reference content, news, premium research, commerce pages, and subscription content may behave differently. If your reporting cannot cleanly isolate a section, document that limitation instead of pretending the test is more precise than it is.

The Publishing Standards rule: maintain an AI Discovery Register

Every professional publisher should maintain a simple AI Discovery Register. It turns a platform switch into an accountable editorial and business decision.

When participation may make sense

Participation can be reasonable when a publication values discovery, brand reach, citations, subscriptions, or new-reader acquisition and can measure whether the additional visibility creates downstream value. Newer sites may also value the opportunity to appear alongside established sources when their content is unusually useful and authoritative.

But “more impressions” is not automatically success. If the publication receives high generative exposure without meaningful visits, signups, leads, sales, citations, or brand growth, the publisher has evidence to reassess the strategy.

When a publisher may consider opting out

A publisher may reasonably consider opting out when its most valuable work is repeatedly summarized without useful downstream value, when a contractual or editorial policy limits participation, or when the publication wants to run a controlled before-and-after test.

The key is to make the choice based on documented objectives rather than fear or hype. AI search is evolving quickly. A setting that makes sense in September 2026 may not make sense six months later.

Do not confuse discoverability with disclosure

A publisher can participate fully in AI search and still need an internal policy explaining how AI was used to create its own content. Those are separate transparency questions. The companion AI content disclosure standard defines a four-level classification for that purpose.

Likewise, discoverability does not end with web search. Ebook accessibility metadata determines whether readers, retailers, and libraries can discover the accessible qualities of a digital book. See the EPUB Accessibility 1.2 release checklist.

FAQ: Google AI search controls for publishers

Can a publisher opt out of Google AI Overviews without leaving normal Google Search?

Google says the new control applies to covered generative Search features and is not used as a ranking signal for search results outside those features. Publishers should still verify their own Search Console settings and ordinary indexing controls.

Is the new control the same as Google-Extended?

No. Google describes Google-Extended as a separate control related to certain uses of content for Gemini model improvement, while the Search Console setting governs participation in generative Search features.

What should a publisher measure before changing the setting?

At minimum: organic traffic, generative impressions, key landing pages, subscriptions, leads or sales, branded search activity, and revenue where relevant.

How often should the policy be reviewed?

Review after a defined test period and again on a regular schedule, such as quarterly, or whenever Google materially changes the feature, reporting, or control.

The standard is the record, not the switch

The lasting publishing lesson is not that Google added another setting. Platforms will continue to change. The durable advantage is having a written decision process that survives the interface.

Professional publishers should be able to answer five questions at any time: What setting did we choose? Who approved it? What evidence did we use? What happened after the change? When will we review it again?

If those answers are documented, AI search becomes manageable publishing infrastructure instead of another trend to chase.

Primary Sources and Further Reading

• Google: New opportunities, control and insights for website owners
• Google Search Central: AI features and your website

The Four Records Every Modern Publisher Needs

This article is part of the September 19 Publishing Standards mini-series. The four records work together: discovery, rights, AI-use transparency, and accessibility.

• AI Training, Piracy, and the Publisher Rights-Provenance Standard
• AI Content Disclosure: A Practical Publishing Standard for 2026
• EPUB Accessibility 1.2: The Publisher Release Checklist

Read the complete September 19 Publishing Standards Brief →