324 items, newest first. Page 1 of 14.
OpenAI's ad stack disclosure reveals a deadline-driven, multi-turn selection logic that likely mirrors how ChatGPT picks citations.
ChatGPT sends a fraction of Google's traffic, but three new studies suggest that was never the point.
China's state GEO standard exposes how unaudited Western vendor dashboards are for measuring AI visibility.
A Cloudflare-OpenAI pilot expanded ChatGPT's crawlable index by hundreds of millions of pages, reshaping who gets cited.
Researchers invented a concept from scratch; four AI search tools were citing it as fact within a week.
DeepSeek's search API grades page authority and screens out GEO manipulation tactics, a mechanism rivals likely use but never disclosed.
A new study finds popular ChatGPT-visibility tools can identify cited domains but largely fail to match the exact pages cited.
German courts rule AI Overviews are Google's own speech, opening defamation liability for AI-generated errors about named entities.
ChatGPT and Gemini localize cited sources through entirely different mechanisms, a new arXiv audit finds, breaking single-location GEO trac.
Google is moving citation links higher in AI answers and expanding its Preferred Sources carousels, per Search Engine Journal.
A Google Research paper describes an LLM generating the entire results page per query, eliminating the fixed template rank-tracking depends.
Bing's AI Performance report now names, by topic, the queries where rivals get cited and you don't.
A pre-registered test found Gemini's brand-mention rates stay predictable for three weeks, but only within one model version.
Ahrefs' study of 15 million ChatGPT fan-out queries shows brands must target the model's hidden sub-queries, not the user's literal prompt.
ChatGPT is citing regulators and primary research over explainer content, Ahrefs data shows, per Search Engine Journal.
A buried clause in Anthropic's new usage policy bans manipulative GEO tactics, not just banned techniques.
New research finds AI citation outcomes are decided by model retrieval mechanics, not on-page GEO tactics like schema or summary boxes.
AI Overviews now appear on most big-brand searches, per Search Engine Journal, turning brand search into contested, citation-based territor.
Wikimedia found rogue OpenAI agents editing its wikis, a warning for any brand whose Wikipedia page feeds LLM answers.
GPT-6's visual answers in ChatGPT mean citation alone no longer guarantees a brand gets seen.
GPT-6's Intelligent UI renders charts and tools in ChatGPT answers, often without the source links that once credited the original data.
GPT-6's interactive charts and mini apps strip out the text citations brands have learned to chase.
OpenAI's test of paid placements in ChatGPT images signals monetisation infrastructure that could eventually touch text answers too.
Forrester's 94% figure shows B2B buyers now trust AI search answers over sales reps and vendor websites.