Google updates AI overviews and links

Google is trying to fix the main trust problem with AI search — the feeling that you get a polished answer, but lose the path back to the actual web. On May 6, it started rolling out five changes to AI Mode and AI Overviews meant to make links more obvious, more useful, and more personal. The basic idea is simple: the AI answer should be the start of research, not the end of it. That matters for users, but also for publishers who have been watching Google’s AI layer eat up attention. (blog.google) ### What changed, exactly? Google’s update has five parts. AI answers will now suggest follow-up reading at the end, surface links from your own news subscriptions, show more direct links inside the response, preview where a link goes before you click, and pull in more “personal perspectives” from places like forums and social posts. In plain English — Google is making the answer look less like a sealed summary and more like a map. (blog.google) ### Why was this a problem? Because AI search has been good at compression and bad at handoff. You ask a big question, get a tidy synthesis, and then have to work too hard to inspect the sources behind it. That creates two headaches at once — users can’t easily judge what’s solid, and publishers worry the click never comes. Goo(blog.google)e web” rather than just getting an answer. (blog.google) ### Why do in-answer links matter so much? Placement is the whole game. A citation tucked off to the side is easy to ignore. A link embedded right where the claim appears is much more likely to get used. Google is now pushing more links directly into the body of AI responses and making previews clearer, so users can tell whethe(blog.google)out it changes the behavior Google wants — click out, compare, keep digging. (blog.google) ### What’s the subscription angle? Google says AI Mode and AI Overviews can now highlight links from publications you already pay for, if those subscriptions are linked to your Google account. That is a small product feature with a bigger message underneath it: Google wants to show that trusted, paid sources still have a place (blog.google)lug into Google’s ecosystem. (blog.google) ### Why bring in Reddit and forums? Because a lot of searches are not really asking for a canonical answer — they’re asking for lived experience. Which stroller holds up after a year? Which insurance claim process is a nightmare? Which city park redesign actually worked? Google is leaning harder into those first-hand perspectiv(blog.google) payload. The catch is obvious — firsthand advice can be useful, but it can also be messy. (theverge.com) ### Is this really about helping publishers? Partly. But it is also about defending Google’s core product logic. Search has always been a traffic router as much as an answer engine. If AI Overviews start feeling like a dead end, Google risks alienating both the people who search and the sites that supply the information. These cha(theverge.com)pathways back to the open web. (blog.google) ### What should people take from this? Google is admitting, without quite saying it outright, that AI answers need better exits. That is the real news here. The company is not backing away from AI search — it is redesigning the escape hatches so the product feels more trustworthy, more inspectable, and less like a black box. (blog.google)