Google Search uses AI to rewrite headlines, sparking accuracy and editorial control concerns

Google Search is now using AI to replace headlines

For example, Google reduced our headline "I used the 'cheat on everything' AI tool and it didn't help me cheat on anything" to just five words: "'Cheat on everything' AI tool." It almost sounds like we're endorsing a product we do not recommend at all. What we are seeing is a "small" and "narrow" experiment, one that's not yet approved for a fuller launch, Google spokespeople Jennifer Kutz, Mallory De Leon, and Ned Adriance tell The Verge. They would not say how "small" that experiment actually is. Over the past few months, multiple Verge staffers have seen examples of headlines that we never wrote appear in Google Search results -- headlines that do not follow our editorial style, and without any indication that Google replaced the words we chose. And Google says it's tweaking how other websites show up in search, too, not just news.

Google Is Using AI to Modify Search Headlines, and It's Not Going Well

Google is reportedly using AI to rewrite some news headlines in Search, and early results indicate the system isn't fully ready yet. As spotted by The Verge, Google is modifying some of the outlet's headlines and occasionally making them incorrect. For example, a headline that was originally written as "I used the 'cheat on everything' AI tool and it didn't help me cheat on anything" was changed to "'Cheat on everything' AI tool." The AI also strips key context from some headlines, making them harder to understand. In one case, it removed the first three words from "I met Olaf -- the Frozen robot who might be the future of Disney Parks," leaving a confusing headline that starts with a lowercase word. These changes are part of an experiment to improve titles on Search and boost "engagement with web content," Google tells The Verge. The latter is contrary to what publishers have been experiencing since the advent of AI Overviews in search. January data from the University of Oxford's Reuters Institute for the Study of Journalism found a steep drop in traffic metrics from 2,756 news sites around the world: 33% worldwide and 38% in the US. Google tested a similar system for Google Discover late last year, before making it a permanent feature for the feed in January. It too has had some accuracy problems. The Verge noticed that Google's AI rewrote the headline on PCMag's story about the FCC foreign drone ban to suggest that the agency had reversed its ban. In reality, our story explains that the Commerce Department ended its drone-ban effort since the FCC had it handled. In early 2025, meanwhile, Apple paused its news-related notification summaries after the feature sent out incorrect blurbs. They returned about six months later.

Google Search 'experiment' uses AI to rewrite news headlines

Chethan is a reporter at Android Police, focusing on the weekend news coverage for the site. He has covered tech for over a decade with multiple publications, including the likes of Times Internet, Guiding Tech, and Android Headlines, to name a few. Chethan's love for Android dates back to the days of the Samsung/Google Nexus S, with his first Android phone being the HTC Desire HD back in 2010. Away from work, he's on the lookout for live cricket streams or NBA highlights. He also enjoys the occasional hour or two of console/mobile gaming whenever time permits. It's no secret that Google is trying to implement AI across practically every product that comes under its vast umbrella. Last year, we saw Google testing AI-generated blurbs for stories in Discover. The company then decided to "experiment" with AI-generated headlines in December, and stopped calling it an experiment over a month later. An extensive report by The Verge has now shed light on a similarly concerning development, but in Google Search. This "experiment" swaps out the original news headline in conventional search results (commonly known as 10 blue links) with AI-generated ones. One of the publication's articles appeared in Search with the following headline: 'Cheat on everything' AI tool. If you hadn't read the article before, you'd think this is an endorsement of the tool. Meanwhile, the original headline read: "I used the 'cheat on everything' AI tool and it didn't help me cheat on anything," which is basically the opposite of what the writer intended to convey (via 9to5Google). Evidently, this can become problematic quickly, particularly as users rely on search engines to find accurate, up-to-date information on current events. The Verge heard back from Google spokespersons about this particular change, who described the experiment's scope as "narrow" and added that a wider launch hasn't been approved yet. Is this just an experiment or a sign of things to come? There are also instances where search results ignore text before or after an em-dash in the headline, as the screenshots above show. The original headlines say the following: "You can't replace the battery in Lego's Smart Bricks -- and many of its sensors aren't active yet," and "I met Olaf -- the Frozen robot who might be the future of Disney Parks." Offering some respite, a spokesperson told The Verge that even if Google went ahead with this experiment, "it would not be using a generative model and we would not be creating headlines with gen AI." However, there was no clarification on what means would be used instead. Another spokesperson said this test is designed to "identify content on a page that would be a useful and relevant title to a users' query." The experiment won't be limited to news articles indexed in Search and will also extend to websites, per Google. It goes without saying that using AI for news headlines is the bigger concern, as such headlines can be read by millions within a few hours of appearing in Google Search. Subscribe to our newsletter for clear AI-in-Search reporting Want clear context on AI headline experiments and their impact on search? Subscribing to our newsletter gives focused coverage and analysis of AI in search and adjacent tech, helping you follow how these experiments change what people read. Get Updates By subscribing, you agree to receive newsletter and marketing emails, and accept our Terms of Use and Privacy Policy. You can unsubscribe anytime. For what it's worth, I haven't seen any instances of Android Police headlines being replaced with AI-generated ones. However, that doesn't necessarily mean it's not happening. There's every chance this test is limited to a few regions, and that Google may soon extend it to other parts of the world. While we hope the idea won't go much further than the experimental stage, considering what happened with Google Discover not long ago, this may be wishful thinking on our part.

Google experiments with AI-generated headlines in search results

Google experiment replaces publisher headlines with AI versions Google is once again testing how artificial intelligence can reshape the way users interact with information online - this time by altering one of the most fundamental elements of search: headlines. The company has begun experimenting with replacing original news headlines in Search with AI-generated alternatives, sparking concerns among publishers and raising questions about accuracy and editorial control. AI starts rewriting the web's "10 Blue Links" In a recent experiment, Google has started displaying AI-generated headlines instead of the original titles written by publishers in its traditional search results. This marks a significant shift from earlier practices where Google might truncate or slightly modify headlines for clarity. Recommended Videos Reports indicate that these rewritten headlines can sometimes change the meaning or tone of the original article. In some cases, nuanced or critical headlines were shortened into generic or misleading phrases, potentially giving users an incorrect impression of the content before they even click. Google describes the feature as a "small and narrow experiment", aimed at improving how well search results match user queries and increasing engagement. However, the company has not disclosed how widely the feature is being tested. Why this raises concerns for publishers Headlines are a critical part of journalism - they are carefully crafted to accurately represent a story while attracting readers. By replacing them with AI-generated versions, Google effectively takes control of how articles are presented to users. Publishers worry this could undermine their editorial voice and credibility. If an AI-generated headline misrepresents a story, readers may blame the publication rather than Google, potentially damaging trust. There are also concerns about transparency Currently, there is little to no indication when a headline has been altered by AI, making it difficult for users to distinguish between original and machine-generated content. This development follows similar experiments in Google Discover, where AI-generated summaries and headlines have already been criticized for inaccuracies and misleading phrasing. For everyday users, the change could subtly alter how information is consumed. Headlines play a major role in shaping first impressions, and even small changes in wording can influence how a story is interpreted. If AI-generated headlines prioritize engagement over accuracy, users may be more likely to click on content that doesn't fully match their expectations. In some cases, they may also miss important context or nuance that was present in the original headline. At the same time, Google argues that AI can help make headlines more relevant to search queries, potentially improving discoverability and making it easier for users to find what they are looking for. What comes next While the current rollout is limited, the experiment could signal a broader shift in how Google integrates AI into Search. The company has a history of introducing features as small tests before expanding them into full-scale products. Publishers and industry observers are likely to push for clearer labeling, greater transparency, and possibly opt-out mechanisms if the feature expands. Meanwhile, regulators may also take a closer look at how AI-generated content is presented, especially when it affects news distribution and public information. For now, the experiment remains a "canary in the coal mine" -- a sign of how AI could increasingly shape not just what users see online, but how that information is framed and understood.

Google Is Using AI to Mess With Headlines (Again)

Google says that if this experiment rolls out into a feature, it won't use generative AI to adjust headlines. Back in December, news outlets spotted Google messing with headlines on articles that appeared in Discover. As it happens, this was Google experimenting with AI-generated headlines; rather than show readers the headlines the outlets wrote themselves, Google's AI would take it upon itself to rewrite them. Why? According to Google, it was a "small UI experiment for a subset of Discover users" that "changes the placement of existing headlines to make topic details easier to digest." While those of us that actually write headlines would argue such an experiment isn't really necessary, Google apparently disagrees, as the experiment is now a feature. The thing is, Google isn't content with keeping AI-generated headlines exclusive to Discover. As spotted by The Verge, the company is now expanding its AI rewrites to Google Search as well. That means if you click an article on a Google Search page, it might have had a headline that the publication had no input on. Of course, when you click through to the actual article, you'll see the real headline. Once again, Google says this is an experiment -- a "small" and "narrow" one at that, and that, in its current state, the feature isn't ready for a full rollout. But The Verge says it has spotted multiple examples of Google taking its articles and changing the headlines for Search. For example, the AI turned the headline "I used the 'cheat on everything' AI tool and it didn't help me cheat on anything" into "'Cheat on everything' AI tool." Sure, the article is about the "Chat on everything" tool, but the AI version leaves out some serious context: The Verge isn't just spotlighting this tool, or, worse yet, endorsing it; it's heavily criticizing it -- something the original headline was designed to advertise to readers. In another example, "You can't replace the battery in Lego's Smart Bricks -- and many of its sensors aren't active yet" was changed to "You can't replace the battery in Lego's Smart Bricks." At least this one doesn't reframe the story, but it does leave out part of the story entirely. Finally, there's "I met Olaf -- the Frozen robot who might be the future of Disney Parks," which Google's AI awkwardly truncated to "the Frozen robot who might be the future of Disney Parks." Yes, the "the" is lowercase in Search, for some reason. The Verge acknowledges that Google has changed far fewer headlines in Search than it has in Discover, which means the chances are much lower that you'll encounter an AI headline in a Google Search than scrolling through the Discover feed. But if the feature's journey in Discover is any indication, Google will only ramp up AI-generated headlines in Search in the near future. Google told The Verge that the goal is to "identify content on a page that would be a useful and relevant title to a users' query," and apply "better matching titles to users' queries and facilitating engagement with web content." This experiment is apparently not specific to news articles, and that, if Google were to launch such a feature, it wouldn't use generative AI to rewrite the headlines. So, why use generative AI to rewrite the headlines in the experiment? To that point, The Verge says that sometimes, Google's AI uses its headline, but the wrong version. Outlets like The Verge will sometimes produce two versions of a headline: one for the website, and one that appears on Search. Google's AI will reportedly swap the headlines, choosing the site headline for Search when The Verge wrote an entire headline designed for Search. Google Search belongs to Google, of course, and we're all at the company's mercy when it comes to what appears in those results. But it seems a bit beyond the pale to change a website's headline when ranking it in Search, whether that be cutting off half the angle, or changing the angle entirely. If readers click through thinking they're getting one thing, and find another, who's that good for? I guess Google thinks it's good for them, but, like many of the changes the company has made in recent years, it isn't really good for anyone else.

Google is now using AI to change search headlines

TL;DR: Google is testing generative AI to alter search result headlines, sometimes changing their original meaning or shortening them significantly. Though currently a limited experiment, Google plans to avoid generative AI in full release, but past AI headline experiments becoming permanent raises concerns about future search accuracy. In a not-so-shocking turn of events, it now looks like Google Search is incorporating generative AI to make questionable changes to the headlines of its search results. Google Search has been a trustworthy search engine for people of all kinds since its inception, with a very straightforward "what you see is what you get" approach. Google CEO Sundar Pichai announced plans to integrate AI into search in 2025, and now it seems they are expanding on it even further. In a recent discovery by The Verge, it seems like Google is putting its own unwelcome twist on several headlines using the power of our "good" friend, generative AI. The Verge uncovered multiple instances in which Google replaced headlines it had written with new ones that sometimes detracted from the original meaning. One particularly egregious example was that the headline "I used the 'cheat on everything' AI tool, and it didn't help me cheat on anything" was changed to "Cheat on everything AI tool." This was particularly infuriating, as it almost entirely changes the meaning the article's author was trying to convey. The new AI-generated headline can easily be mistaken for a product endorsement rather than the scathing criticism it was. Google spokespeople responded to a query from The Verge, stating that the changes were part of a "small" and "narrow" experiment that had not yet been approved for a full launch. Google's words do not fill us with confidence that this practice will not become mainstream in the near future. Google claims that the new dynamically generated headlines are based on user context, search history, location, and other personalized factors. This means that the same webpage can be displayed with different headlines depending on who's typing the search query. According to Google spokesperson Mallory De Leon, the test currently uses generative AI, but the company does not plan to use it for the full release. "If we were to actually launch something based on this experiment, it would not be using a generative model, and we would not be creating headlines with generative AI." - De Leon responded to The Verge Now, a sliver of good news is that these headlines seem uncommon and not that egregious in the few examples we have seen. Mostly, Google just seems to be shortening every headline into a digestible 4-to-5-word package. However, it is fair to say that the short, boring headlines do not convey the full meaning the publisher intended. Google is firm on its stance that this is a very limited experiment; however, it also claimed that the AI-generated headlines found in Google Discover earlier this year were an experiment. They ended up making the feature permanent after all, which does not bode well for the future of Google Search.

Google Tests AI Headline Rewrites in Search Results

Google is using AI to rewrite article headlines and page titles in Search results, the company confirmed to The Verge, per Search Engine Land, a move that publishers cannot opt out of. This follows a pattern, as Google has made a similar experiment permanent before. Did publishers agree to this? Google's Search Central documentation on title links, published since 2021, states that Search titles are auto-generated from any text on a page: headings, meta tags, and anchor text. Most publishers did not know this included AI-generated rewrites. By remaining indexed on Google Search, publishers accept terms that permit this. Google offers no opt-out mechanism for individual policies and no way to flag objections to specific rewrites. It is all or nothing. Has Google done this before? Google tested AI headline rewrites in Discover, its personalised news feed, describing it in December as a limited experiment. Within a month, Google made that test a permanent feature, citing strong user satisfaction scores as justification. Publishers expressed concerns at the time, but Google did not address them. The same experiment is now running in traditional Search, which carries higher stakes because Search remains the primary discovery channel for most news content. While Discover is opt-in and feed-based, Search is not. Regulators elsewhere are beginning to push back. European publishers have filed an antitrust complaint against Google over AI Overviews, and Google is developing opt-out controls for AI search features under pressure from the UK's Competition and Markets Authority (CMA), though MediaNama founder Nikhil Pahwa has argued this amounts to a "facade of choice" since opting out risks disappearing from AI-generated answers altogether. What is Google actually testing? The test is "small" and "narrow," and Google has not approved it for a broader rollout. It affects news sites, but is not limited to them. If Google does roll it out widely, the company says it would not use a generative AI model, meaning the system would select and trim existing text from a page rather than compose new headline text from scratch. Google has not explained how that would work in practice or whether it would produce results that are meaningfully different. What are publishers saying? Sean Hollister, senior editor at The Verge, compared it to "a bookstore ripping the covers off the books it puts on display and changing their titles." Louisa Frahm, SEO director at ESPN, wrote on LinkedIn that if headline facts get misrepresented, long-term audience trust will suffer. The concern is not abstract: Google reduced a headline reading "I used the 'cheat on everything' AI tool, and it didn't help me cheat on anything" to "'Cheat on everything' AI tool," stripping the article's entire conclusion. It changed another, "Microsoft is rebranding Copilot in the most Microsoft way possible," to "Copilot Changes: Marketing Teams at it Again." Why does this matter when traffic is already falling? AI chatbots and Google's AI products account for less than 1% of publisher page view referrals, per a 2025 industry report, and a Pew Research Center study found that users click on links within AI Overviews only 1% of the time. Headline rewrites mean publishers are now losing not just traffic volume but control over how their remaining Search presence looks. What does this mean for Indian publishers? India has no regulatory framework that gives publishers formal leverage over how platforms display their content. Australia's News Media Bargaining Code and Canada's Online News Act both require platforms like Google to negotiate compensation and display terms with publishers, giving them a seat at the table. India has no equivalent, meaning Indian publishers have no formal mechanism to push back against Google's changes to how their content appears in Search. The linguistic risk compounds this. Most general-purpose AI models demonstrate lower accuracy in Indian languages than in English, particularly beyond Hindi and Tamil. A headline rewrite in Marathi, Kannada, or Bengali carries a higher risk of factual distortion or loss of culturally specific framing. Indian-language publishers also lack the SEO infrastructure to systematically monitor discrepancies between what they publish and what Google displays, so inaccurate rewrites may go undetected and uncorrected.