Google is quietly changing how AI-powered search works by letting you bring your own trusted sources into the mix – and that simple tweak could end up being one of the most important updates for publishers, creators, and newsrooms since AI Overviews first showed up in the results page. With Preferred Sources now rolling out directly inside AI Overviews and AI Mode, your readers can effectively “star” your site and see it called out with a visible badge whenever Google’s AI leans on your work to answer a query.

For anyone who makes a living from content, this is the first time the AI layer in Search has given users a formal way to say “I want more from this publication – and I want that preference to matter when you summarize the web for me.”

If you haven’t followed every twist in Google’s AI search story, here’s the short version: over the past couple of years, Google has been moving toward a world where many queries are answered first by an AI-generated overview, with classic blue links and Top Stories pushed further down the page. AI Overviews condense information from multiple pages into a single block, and AI Mode takes that even further, turning Search into something closer to a chat-style assistant that talks back and refines answers as you ask follow-up questions.

That shift has made a lot of publishers and creators nervous, because when AI is summarizing the web for you, there’s always a risk the user stops at the summary and never clicks through to the sites that actually did the work. Multiple reports and industry commentary have pointed out that AI answers, by design, sit above traditional links and can absorb a big share of user attention. In that context, anything that better connects AI answers back to original sources is a very big deal.

Preferred Sources started life as a more modest personalization feature: you could pick your favorite outlets in Top Stories, and Google would try to surface them more often for news-heavy queries. Earlier this year, Google expanded that model with a dedicated “source preferences” area inside Search personalization settings, where you can search for sites you like and mark them as preferred so they show up more prominently in news carousels and content-focused search features. Now, with this rollout, the same preference finally extends into AI Overviews and AI Mode – meaning your choices influence not just which links appear, but which links get highlighted inside AI-generated answers.

So what actually happens when Preferred Sources meet AI Overviews?

From the user’s side, the experience is pretty simple. You go into Search personalization, find “Source preferences,” and start adding websites you trust – any site that regularly publishes fresh content is eligible, whether it’s a big news brand, a niche blog, or a specialist newsletter you rely on. Once that’s set up, whenever you run a query that triggers an AI Overview or you switch into AI Mode, links from your preferred outlets have a better chance of being pulled in as supporting citations, and when they do appear, they’re labeled with a clear “Preferred” badge next to the link that only you can see.

Under the hood, Preferred Sources is no longer just a cosmetic tag. Google has confirmed that these preferences are used as an actual signal when the system decides which publishers to include in AI responses, as long as the content is relevant and up to date. In other words, if you and I both search the same topic and we’ve chosen different favorite sites, the AI Overview we see could be grounded in slightly different supporting links, even if the generated text looks similar. That’s a subtle but significant shift away from one-size-fits-all AI answers and towards something more personalized and relationship-based.

Google is also layering in a new carousel treatment in AI Mode that leans on Preferred Sources, making it easier to jump directly into articles and coverage from outlets you’ve already signaled you like. Combined with the badge, it effectively turns AI Mode into a sort of smart reading hub where your trusted brands get front-row placement in the AI-powered interface, rather than being buried under generic aggregated results.

The other big piece of context here is Google’s broader push around “original, high-quality content” and signals like the “Highly Cited” badge that highlight where a story or key piece of reporting started. Over the last couple of years, Google’s core updates have repeatedly emphasized usefulness, expertise, and real-world experience, while cracking down on thin, low-value, or spammy AI-driven pages. The company’s latest explainer on how Search surfaces original reporting explicitly calls out both Highly Cited labels and Preferred Sources as tools designed to help users discover authoritative, first-hand coverage more easily.

Bringing those philosophies into AI Overviews and AI Mode is important because it signals that Google doesn’t want its AI answers to float in a vacuum. Instead, the company is trying to position AI as a layer that still routes people back to original work, especially when that work has earned trust through consistent quality or a direct relationship with readers. For publishers, that’s at least an acknowledgement that Search needs to remain a two-way street: Google gets content to train models and answer questions, but the sites it draws from still need traffic, visibility, and loyal audiences to survive.

If you are on the publishing side of this equation, Preferred Sources inside AI Overviews is both a traffic opportunity and a new to-do list item. Google and several SEO-focused observers have already pointed out that users are significantly more likely to click on a Preferred Source link inside AI experiences than on a generic citation. Some analysis shared by search marketing firms suggests people may be roughly twice as likely to click when they recognize the brand and see that “Preferred” label, which turns a standard citation into a kind of personalized endorsement.

Unsurprisingly, Google is nudging site owners to actively campaign for that status. The company’s documentation describes a dedicated deep link format – essentially a URL that opens directly to your site’s entry in the source preferences tool, where users can tick a single box to add you as a preferred outlet. There are even downloadable “Add as preferred source” buttons you can drop next to your social follow icons, suggesting Google anticipates publishers treating this like another must-have call to action, right alongside “Subscribe” and “Follow us on X.”

In practice, that means we’re likely to see a new pattern emerge in newsletters, social posts, and even on-site marketing: “Add us as a Preferred Source on Google.” Think of it as the search-era equivalent of asking readers to ring the notification bell on YouTube – a small personalization tap that can meaningfully change how often you show up in the feed, or in this case, in AI’s field of vision.

From a user standpoint, Preferred Sources inside AI Overviews is about control. Search has always been a mix of algorithmic ranking and personal habit, but traditional ranking signals – links, authority, freshness, engagement – don’t always line up with individual trust. You might personally rely on a smaller specialist blog for camera reviews, or a local newsroom for city politics, even if bigger outlets outrank them in raw SEO terms.

By letting you put a thumb on the scale and saying, “If you’re going to summarize this topic for me, please include this site when it’s relevant,” Google is quietly acknowledging that expertise is relational, not just statistical. That’s especially important in an era where AI answers feel more abstract and less obviously connected to the underlying sources. Seeing your trusted titles called out – and knowing the system is actively trying to include them – can rebuild some of the transparency that was lost when everything moved into a single AI block.

There is also a small but meaningful personalization dimension for news consumption. If you’ve marked your preferred outlets, your AI Overviews for fast-moving stories could start to feel less like anonymous wire copy and more like a curated digest that reflects where you usually read. The text you see may still be generated by an AI model, but the underlying links – and the direction you’re nudged to click – are shaped by your own sense of who you trust.

Of course, there are caveats. Preferred Sources only matter when your chosen outlets are publishing fresh, relevant content on the query you asked – the system isn’t going to shove in your favorite cooking blog as a source on a complex medical query just because you starred it once. Google’s documentation stresses that the usual technical requirements still apply: your pages need to be indexable, meet Search policies, and be eligible to show a normal snippet before they can ever show up as links in AI Overviews or AI Mode.

There’s also the broader question of how much personalization is too much. If AI answers become increasingly tailored to whichever outlets a user prefers, there’s a risk of tightening information bubbles, where people mostly see summaries anchored in sources they already agree with. On the flip side, that’s arguably not very different from how many of us already use Search and social – we gravitate toward familiar brands and voices – and at least in this case, AI is surfacing multiple citations rather than just one opinionated post. The real test will be whether Google manages to keep a baseline of diversity and quality while still honoring individual preferences.

From a competition standpoint, Preferred Sources could tilt the playing field in favor of brands that already have strong direct relationships with their audience. If you’re a well-known publication with a big newsletter list, you can ask hundreds of thousands of subscribers to add you as a preferred source in a single email blast. Smaller or newer outlets may struggle to secure the same level of reader participation, especially if they’re still working on basic discovery and brand recognition. But for niche sites with loyal communities, this is also a rare chance to translate that loyalty into a concrete search advantage.

The interesting question is what this move says about Google’s roadmap for AI in Search. Over the last year, the company has been steadily rolling advanced model capabilities into the core search flow, from multi-step reasoning to “agentic” behaviors where AI doesn’t just answer questions but helps you take actions based on them. AI Mode sits right at the heart of that ambition – a dedicated environment where you can treat Search like an assistant, ask for planning help, and iterate with follow-ups instead of firing off one-off queries.

By weaving Preferred Sources into both AI Overviews and AI Mode, Google is effectively acknowledging that even the most capable AI system needs grounding in human editorial judgment – not just at model training time, but at the moment of retrieval. It’s an admission that “AI plus the open web” works best when users have some say in which corners of the web the AI should prioritize, and when the system makes those choices visible through labels and carousels rather than hiding them inside a black box.

For creators and publishers, the message is equally clear: AI search is not going away, so the question now is how to plug into it strategically instead of just bracing for impact. Being selected as a Preferred Source doesn’t guarantee you traffic, but it does give you a recognizable foothold inside the AI surfaces that increasingly sit between your content and your audience. Combined with ongoing pushes around original reporting, Highly Cited labels, and people-first content, Google is sketching out a future where the sites that win are those that both satisfy algorithmic quality thresholds and build enough trust with real readers that people are willing to actively prioritize them.

In that sense, Preferred Sources may end up being the most human part of Google’s AI search evolution: a simple little checkbox that quietly decides which voices get carried through into the AI-generated future of Search.