Publisher opt-out from AI search: Google is testing a setting that allows publishers to exclude their pages from AI search summaries, AI mode, and Discover, while keeping them in traditional search results.
Article index
- New control over smart search features
- The option does not equal blocking Gemini training
- British trial before global rollout
- No announced penalty for traditional search ranking
- Additional data to help publishers
- Frequently asked questions
New control over smart search features
Google is working on a Search Console setting that allows publishers to exclude their content from AI summaries, AI mode, and generative features in Discover, while keeping their pages appearing in traditional search results. The step comes following requirements from the UK Competition and Markets Authority to give publishers clearer control.
The option does not equal blocking Gemini training
The announced control pertains to content visibility and usage within generative search features, and should not be automatically described as a blanket ban on using articles to train Gemini models. Training, crawling for indexing purposes, and citing content within a generated answer are distinct processes that require separate controls and definitions.
British trial before global rollout
Google stated it will begin testing with a group of publishers in the UK and then roll out the setting globally. When opting out, a publisher’s pages will not appear as links within generative features and will not be used to support answers in them, while the crawling necessary for regular search remains active according to other site settings.
No announced penalty for traditional search ranking
The company confirmed that the opt-out decision will not be used as a signal for page ranking in traditional search. However, publishers may lose visibility and traffic coming from AI summaries or AI mode, meaning the decision involves a trade-off between protecting usage within these features and benefiting from their growing presence.
Additional data to help publishers
Google plans to provide metrics in Search Console regarding page impressions within smart answers and the countries where those impressions occurred. This data can help media organizations gauge the impact of the features before opting out, though publishers are still demanding more precise details on clicks, citations, and commercial returns.
This means that publishers opting out of AI search is not a simple technical decision. An organization needs to compare the traffic and referrals it receives from the new features against the value of the content Google uses in its answers. The setting should also be tested on a subset of pages and Search Console data monitored before rolling it out across a large site.
Returning to original sources helps refine the meaning of numbers and distinguish between an official announcement, a press report, or an analytical estimate. Such comparisons also prevent future plans from being turned into accomplished facts before measurable operational evidence appears.
Impact assessment is not limited to investment volume or announcement strength, but includes actual performance, cost, scope of use, and human oversight. These indicators provide a more accurate picture than marketing comparisons alone.
Some details may change with the release of technical documents, regulatory statements, or subsequent independent results. Therefore, monitoring remains necessary, especially when numbers are attributed to a company or an initial report.
The broader impact depends on the relevant entity’s ability to turn the announcement into a stable product or service. Critical tests include execution reliability, data protection, and clarity of accountability in the event of an error.
From the user’s perspective, value appears when access improves, costs decrease, or quality rises noticeably. From the market perspective, sustainable adoption remains more important than temporary reach.
A distinction must be made between this option and preventing models from training on content. The proposal relates to the appearance of pages within Google’s generative search experiences, while crawlers used for training or other services are subject to separate controls. This difference is important for publishers when balancing traditional traffic with the desire to limit content paraphrasing inside answers.
Sources: TechCrunch, Search Engine Journal
Frequently asked questions
Question: What can a publisher exclude?
Answer: AI summaries, AI mode, and generative features in Discover.
Question: Does the page disappear from regular Google search?
Answer: No, Google says exclusion does not remove pages from traditional search results nor is it used as a ranking signal.
Question: Does the setting prevent Gemini training?
Answer: The announced setting concerns generative search features, and does not alone prove a comprehensive ban on model training.