Why AI Overviews Favor Newer Content Over Established Authority

AI assistants do lean toward newer content, but in the largest study of the question Google’s AI Overviews went the other way. Ahrefs analyzed nearly 17 million cited URLs and found that content cited by ChatGPT, Perplexity, Copilot and Gemini was, on average, 25.7% fresher than organic Google results, with ChatGPT showing the strongest preference for new pages. The top three citations in Google’s AI Overviews were, on average, 16 days older than the organic results. So where an AI Overview does cite a newer page over an established one, a more likely explanation is how Google’s ranking systems treat freshness for that query, not a special preference built into AI Overviews.

Freshness is a query-by-query decision in Google

Google’s guide to its ranking systems describes “query deserves freshness” systems “designed to show fresher content for queries where it would be expected.” Its examples make the logic clear: someone searching for a newly released movie probably wants recent reviews, and a search for “earthquake” might normally return preparation material but surface news if an earthquake has just happened.

AI Overviews draw on those same systems. Google’s AI optimization guide says AI responses are grounded by “relying on our core Search ranking systems to retrieve relevant, up-to-date web pages from our Search index.” When a query is one where freshness matters, the retrieved pages tend to be newer, and the AI Overview built from them reflects that. When freshness does not matter, established pages compete on their usual strengths.

There is one notable exception at the extreme end. Google has said it holds back AI Overviews on the most time-sensitive topics. In a May 2024 post on AI Overview errors, Liz Reid of Google wrote that “we aim to not show AI Overviews for hard news topics, where freshness and factuality are important.” For breaking news, the freshest content may not appear in an AI Overview at all, because the overview may not appear.

What the measurement shows across assistants

The Ahrefs freshness study, published in July 2025, compared the age of cited URLs on each platform with organic results:

• Across ChatGPT, Perplexity, Copilot and Gemini, cited content was 25.7% fresher than organic Google results.
• ChatGPT showed the strongest preference, citing URLs 393 to 458 days newer than organic Google results.
• Google’s AI Overviews were the exception, citing content that was 16 days older than organic results on average.

The comparison suggests that the “AI favors new content” effect is real but uneven. In that study it was strongest for ChatGPT and reversed for Google’s AI Overviews, which sit on top of Google’s established ranking.

When newer content does win in AI Overviews

A newer page can be cited ahead of an established one for reasons that have little to do with age itself.

The query deserves freshness. Topics that change, such as software features, prices, regulations, product lines and current events below the hard-news threshold, are the kind of queries Google’s freshness systems are designed for. A recently updated page has an advantage there.

The established page is out of date. An authoritative page that has not been updated may contain facts that are no longer true. A newer page with current information answers the question better, regardless of the older site’s reputation.

The newer page answers more specifically. If an established page covers a topic broadly and a newer page addresses the exact question, the newer page may be the better match for the passage the overview needs.

Fan-out reaches narrower questions. Google documents a “query fan-out” technique in which AI Overviews can draw on related searches across subtopics. A newer page focused on a subtopic could be retrieved this way. Ahrefs found little support for fan-out in its 2025 study but read its 2026 data as pointing toward it, and Google does not say how often it happens, so treat it as a possibility rather than a rule.

What established sites should do

1. Identify pages on topics that change. Prices, features, rules and statistics go stale. Those pages are where newer competitors can overtake you.
2. Update with substance, not just dates. Changing a timestamp without changing the content does not make a page more current. Add new information, remove what is no longer true, and then show the update date.
3. Keep evergreen pages accurate. For topics where freshness does not matter, established pages compete on depth and trust. Check them for errors rather than rewriting them for the sake of novelty.
4. Answer specific questions inside authoritative pages. Make sure your strongest pages state current, specific answers clearly, so they remain the best match for the passages an overview draws on.
5. Watch other assistants separately. If ChatGPT visibility matters to you, recency weighs more there, according to the Ahrefs data, and regular updates matter more.

Treat freshness as a property of the query, not a bias of AI Overviews. Where Google’s systems expect current information, update your pages with real changes; where they do not, keep established pages accurate and specific. In Google’s AI Overviews, established authority still counts, provided the content behind it is current.