ChatGPT Tests Web-Only Search Results and Google-Style 'See More' Answers
OpenAI is testing web-only results, reordered answer cards and a Google-style "See more" collapsed AI answer in ChatGPT search, per three X sightings.

- OpenAI is testing at least three ChatGPT search result formats, all spotted within the past week
- One test shows a collapsed AI answer with a "See more" button above web results, mirroring Google's AI Overviews layout
- Sightings came from Joe Youngblood, Juozas Kaziukėnas and Fernando on X; the reporter could not replicate any of them
- The web-only results view had been seen before and was initially dismissed as a bug
- OpenAI has not confirmed or announced any of the tested formats
OpenAI is running several simultaneous tests of how ChatGPT presents search results, including a web-only results format, a reordering of its richer answer cards, and a collapsed AI answer with a "See more" button that mirrors the pattern Google uses for AI Overviews (Google's AI-generated answer summaries shown above traditional links). All three variations surfaced within the past week through independent sightings on X.
Barry Schwartz, who documented the tests, noted plainly: "I couldn't replicate any of these findings myself." That caveat matters. These are unconfirmed interface experiments, not announced product changes — OpenAI has made no public statement about any of them.
What exactly did testers spot?
Three distinct formats appeared in screenshots shared on X:
- Web-only results: Joe Youngblood spotted ChatGPT returning a plain list of web results, without the AI-generated answer cards the interface normally produces.
- Web results first, richer results second: Juozas Kaziukėnas observed ChatGPT displaying web results first and then following with its normal, more elaborate answer format — an inversion of the current default order.
- AI answer with a "See more" button: Fernando captured ChatGPT showing the AI answer at the top with a "See more" expansion button, followed by web results underneath.
The third test is the most structurally significant. It replicates the core layout logic of Google's AI Overviews, where a condensed AI summary sits above organic results and expands only when the user asks for more. If OpenAI ships that pattern, ChatGPT would treat the AI answer as the primary result and the web listings as supporting material — the same hierarchy shift publishers have been absorbing in Google Search since AI Overviews rolled out broadly in 2024.
Confirmed versus speculative
What is confirmed here is narrow: screenshots from three separate users, published on X, showing three different interface behaviors in the same week. Schwartz's coverage stands on those observations.
What remains speculative is intent and rollout. A/B tests in ChatGPT have appeared and disappeared before, and one of the formats — the web-only results view — had been seen previously and was initially dismissed as a bug, Schwartz wrote, explaining why he had not covered it earlier. It only earned coverage now because it fits a pattern of multiple search-related experiments landing in the same week.
No deployment dates, traffic figures or publisher impact data exist for any of these tests. Nobody outside OpenAI knows whether any format will reach general availability.
What does this change for site owners and SEO professionals?
In practical terms, nothing yet. But the direction of travel is legible. OpenAI keeps refining how ChatGPT cites and lists web sources, and each experiment moves the product closer to a conventional search engine — with ranked links, an answer layer, and a defined visual relationship between the two.
For publishers, the "See more" test deserves the closest attention. A collapsed AI answer above a web results list is the exact configuration that reshaped click-through behavior on Google. If ChatGPT adopts it permanently, sites already tracking referrals from ChatGPT search would likely see the same bifurcation Google's results produced: visibility inside the AI answer, with clicks concentrated among the few links shown beneath it.
The web-first ordering test cuts the other way. Placing web results above the richer AI cards would push users toward source sites sooner — a configuration publishers would welcome and one that would differentiate ChatGPT from Google's answer-first approach.
What to watch next
The signal to monitor is whether any of these three formats survives testing and reaches all ChatGPT users, and in what order OpenAI settles the hierarchy between its AI answer and the web listings beneath it. As Schwartz signed off: "Always be testing."
via x.com (Original)
More from Tom Whitfield
Show full bio
News editor covering regulation, policy and their impact on Search rankings and algorithm updates, generative engine optimization at SERP Journal.
183 articles