Google is testing a feature that replaces publisher-authored headlines in standard Search results with AI-generated alternatives, The Verge reported after documenting multiple rewrites to its own articles.
The experiment sources replacement text from content already present on the page rather than generating new titles from scratch. Google representatives confirmed the test is running but described it as "small" and "narrow," with no broader rollout yet approved. There is currently no opt-out mechanism for publishers.
What the headline rewrites look like
The Verge documented several instances in which its headlines were rewritten without notification or attribution. In one case, an article titled "I used the 'cheat on everything' AI tool and it didn't help me cheat on anything" appeared in Search as "'Cheat on everything' AI tool," removing the article's critical editorial stance entirely. The rewrite preserved factual information while discarding the journalist's interpretation.
Google clarified that the system does not use generative AI to invent new titles. Instead, it draws on text already present on the page, such as subheadings, pull quotes, or body copy, and surfaces it as an alternative to the headline. Whether that distinction matters to publishers depends on how closely the substitute text represents the article's angle.
The News Media Alliance, which represents thousands of U.S. news outlets, has called on Google to provide transparency about when rewrites occur and to create an opt-out mechanism for publishers.
Why traditional Search raises the stakes
Google has previously experimented with headline modifications in Discover, its algorithmically curated content feed. This test extends the practice into traditional Search results, which remain the primary discovery channel for most news content online.
The distinction is significant for publishers with SEO-dependent traffic models. Title tags and headlines are among the most carefully optimized elements of a news article: they affect click-through rates directly, shape how readers understand the story before clicking, and often reflect deliberate editorial decisions about framing. A replacement drawn from body copy may be factually accurate but misrepresent the editorial stance the publication chose to take.
Louisa Frahm, SEO Director at ESPN, noted that misrepresented headlines could damage long-term audience trust, an effect that would not appear in standard traffic metrics.
The broader pattern of editorial override
The headline rewriting test is the latest step in a gradual shift in how Google treats publisher-authored metadata. AI Overviews already summarize content without requiring users to click through. Google has also modified title tags in results when it judges them to be low quality or misaligned with the page's content. The headline experiment extends this logic further, applying a subjective editorial judgment to the most visible element of a search result.
Microsoft Bing has also been modifying title tags as part of its AI-powered search rollout, suggesting the practice may become standard across major search engines.
For publishers, the compounding effect is a narrowing of editorial control at every layer of how their journalism appears in search. AI Overviews compress or summarize the story; headline rewrites change the framing; both occur without consent and without recourse. Penske Media sued Google over AI Overviews in 2025, naming traffic loss and editorial control as core concerns.
Sean Hollister at The Verge described the situation as "like a bookstore ripping the covers off the books it puts on display and changing their titles." The analogy captures the core issue: the change is not about access to content, but about who controls how it is presented.
What publishers and SEO teams should do now
There is currently no documented way to opt out or flag objections to specific rewrites. Publishers should audit Search performance for articles where click-through rates have dropped unexpectedly, as headline rewrites are one possible explanation for underperformance that would not appear in standard analytics.
Google's existing publisher controls documentation, which covers opt-outs for AI features broadly, does not address headline rewrites specifically.
Recap
Can publishers opt out of Google rewriting their headlines in Search?
Not currently. Google confirmed the test is running but has not provided any opt-out mechanism for publishers who object to specific rewrites. The News Media Alliance has called on Google for transparency and opt-out tools, but no public response has been published.
How does Google's headline rewriting affect SEO and click-through rates?
Google is sourcing replacement text from content already on the page, not generating new copy. If the substitute text is less descriptive or strips editorial framing, it can reduce CTR in ways outside publishers' control, making headline optimization and A/B testing less reliable for Search performance.
Is this the same as Google's existing practice of rewriting title tags?
This goes further. Google already sometimes substitutes title tag text with an H1 or other on-page element. This experiment applies similar logic specifically to headline rewriting in both traditional Search results and Discover, with AI selecting replacement text from the page rather than the publisher's chosen title.
