Does independent evidence support Carragher & Carley's modeled estimate that a full shift to Google AI Overviews would raise US media concentration ~3.4x more than the Paramount–WBD merger did?
This capture is a targeted follow-up to question-corroborate-carragher-carley-2026-ai-overview-concentration-scenario, raised during the 2026-10-31 promotion of arXiv:2610.40164 (the same paper behind claim-owner-level-news-concentration-130x-outlet-level and claim-ai-citation-concentration-exceeds-traffic-concentration). The question asks whether any source other than Carragher & Carley's own preprint corroborates — or disputes — their modeled estimate that a full shift of US organic search to Google AI Overviews would raise US media concentration by roughly 3.4x the effect of the Paramount Skydance–Warner Bros. Discovery merger.
A genuine search effort (arXiv, general web, trade press, academic-seminar listings) found: the primary estimate itself (restated below for grounding); no independent replication or rebuttal of that specific multiplier; and one independently authored causal study that corroborates the underlying mechanism the scenario assumes, without testing the scenario's concentration math. The central question is marked unverified below — this is reported as a complete, honest outcome, not a gap to keep chasing.
Claim: Carragher & Carley's own modeled estimate — a full shift to Google AI Overviews would raise US media concentration 3.4x more than the Paramount–WBD merger did — is a stated counterfactual scenario, not a measured historical effect
verifies: question-corroborate-carragher-carley-2026-ai-overview-concentration-scenario
The paper takes the actual Paramount Skydance acquisition of Warner Bros. Discovery (closed 6 October 2026, contested by twelve state attorneys general) as an empirical baseline: combining the two companies' news outlets under one owner raises the authors' traffic-based estimate of US media concentration by +9.9%. Against that baseline, the authors construct a counterfactual in which all US organic search traffic is redistributed in proportion to Google AI Overviews' current citation pattern (holding total traffic constant, and starting from an assumption that AI Overviews presently mediate "a negligible share" of news traffic). Under that counterfactual, their concentration estimate rises by +33.3% — about 3.4 times the merger's measured effect. The paper's own text flags this as a scenario built on an assumption about how traffic would redistribute, not an observed shift: confirming that traffic actually follows AI citation patterns over time "requires longitudinal data" the authors say they do not have.
- source_url: https://arxiv.org/abs/2610.40164 (PDF: https://arxiv.org/pdf/2610.40164)
- source_sha: 33521fd5dc14fbd653efcf1482cf3bbd35bd9ea16d212128846c8b0cea92e1f6
- source_author: Peter Carragher; Kathleen M. Carley
- source_date: 2026-10-29
- source_title: "Who Owns Digital News, and Who Gets Cited: Measuring Media Concentration Across 77 Countries and Seven AI Systems"
- source_venue: arXiv:2610.40164v1 [cs.CY]
- source_tier: 1
- source_quote: "Combining the two groups' outlets under one owner raises our traffic-based estimate of US media concentration by +9.9%. Now take the current market as the baseline, in which we assume that AI overviews mediate a negligible share of news traffic, and consider the case where all organic search traffic passes through Google AI Overviews. In this scenario, with attention redistributed in proportion to Google AI Overview's citations, US media concentration rises by +33.3%, roughly 3.4 times the Paramount merger's effect."
- claim type: quantitative + counterfactual/modeled — Tier 1 for the authors' own stated figure, but explicitly their own scenario projection, not an observed outcome.
Claim: No independently authored study has replicated, tested, or disputed the specific 3.4x / +33.3% concentration-multiplier estimate
verifies: question-corroborate-carragher-carley-2026-ai-overview-concentration-scenario
Searches for a second author group applying comparable ownership/HHI-concentration methodology to an AI-Overviews-redistribution scenario — on arXiv, in general web search, in media-industry trade press (Press Gazette, Hacks/Hackers, ProMarket, MLex), and in academic seminar listings covering AI-and-publisher economics — surfaced no independent attempt to reproduce, extend, or challenge this specific figure. Coverage of the paper itself (e.g. the Hacks/Hackers write-up of the same preprint) restates the authors' own number rather than testing it against a second dataset or model. No contemporaneous critique of the counterfactual's assumptions (e.g., holding total traffic constant, assuming AI Overviews presently carry "a negligible share" of news traffic) was found either. As of this search (2026-11-01, two days after the preprint's submission), arXiv:2610.40164 has not been confirmed as accepted to any peer-reviewed venue, though its own text frames itself as a submission to a "Web Economics and Digital Society" conference track.
This claim is a negative/absence finding about the state of the literature rather than a quotation from a single document; it rests on the search process described above, not on one source. Per the sourcing floor's handling of genuine search failure, the core question is marked here: [unverified — could not confirm or deny after search].
Claim: An independently authored causal study corroborates the underlying mechanism the scenario assumes — that AI-generated search answers divert traffic from publishers — without testing the concentration-multiplier magnitude itself
A different author group, at a different institution and venue, provides causal (not merely correlational or modeled) evidence that Google AI Overviews measurably redirect search traffic away from the publishers that supply it. Mehrzad Khosravi and Hema Yoganarasimhan (University of Washington) use AI Overviews' staggered geographic rollout together with Wikipedia's multilingual structure as a natural experiment: a difference-in-differences design compares monthly external-search referrals to English Wikipedia articles against referrals to the same articles in German and French. Default AIO availability reduced English search traffic by 5.45% (vs. German) and 4.82% (vs. French). This is independent in the sense the vault's single-source cap requires — a different author group and a different venue from Carragher & Carley — and it supports the plausibility of the mechanism their AI-Overviews scenario assumes (that AI-mediated answers reallocate search attention away from source pages). It does not, however, use concentration/HHI methodology, does not model ownership-level market structure, and does not test or confirm the 3.4x multiplier itself — it corroborates a precondition for the scenario, not the scenario's magnitude.
- source_url: https://arxiv.org/abs/2602.18455
- source_sha: 7c9cc03a8f5c699909df914c4e1d7b44d0e1d51fa15943c0e03707805234a12a
- source_author: Mehrzad Khosravi; Hema Yoganarasimhan
- source_date: 2026-10-25 (arXiv v7; article text dated "October 27, 2026")
- source_title: "Impact of AI Search Summaries on Website Traffic: Evidence from Google AI Overviews and Wikipedia"
- source_venue: arXiv:2602.18455v7 [cs.CY]
- source_tier: 1
- source_quote: "finds that default AIO availability reduced English search traffic by 5.45% and 4.82% , respectively. Our results suggest that answer-producing digital intermediaries can materially reallocate attention away from informational publishers, with implications for content monetization, search platform design, and policy."
- claim type: quantitative + technical-mechanism — Tier 1 (own causal diff-in-differences estimate on the authors' own data) — clears the floor for both.
Further leads
- Saharsh Agarwal (Indian School of Business) and Ananya Sen (Carnegie Mellon, Heinz College), "The Impact of Google AI Overviews on Publisher Traffic and User Experience: Evidence from a Field Experiment" — a pre-registered field experiment (~1,000+ US Chrome users) reportedly finding AI Overviews cut organic outbound clicks by ~39.8% and raised zero-click searches by roughly a third; read here only via ISB's own institutional research summary (https://www.isb.edu/faculty-and-research/isb-institute-of-data-science/research/ai-summaries-in-search-and-their-impact-on-publisher-traffic), not the primary SSRN/arXiv paper —
[unverified-quant — needs primary]. - A Graphite/Similarweb study of 40,000+ large US websites, and separate Reuters Institute/Chartbeat figures reportedly showing a ~33% global decline in Google search traffic to publishers, surfaced in secondary coverage only — not independently fetched or verified this session.
- Whether arXiv:2610.40164 has since been accepted to a refereed venue (the paper's own text frames itself as targeting a "Web Economics and Digital Society" track) is worth periodic re-checking.
- Dominic Ponsford / Press Gazette trade coverage of AI-citation concentration (cited as ref [37] in Carragher & Carley's own paper) remains unfetched independently.
Entity candidates
- Eli M. Noam — person (media economist, Columbia University) — concept/term — author of Who Owns the World's Media? (2016), the foundational prior work Carragher & Carley's whole ownership-concentration methodology is checked against (already flagged in the 2026-10-31 capture; repeating here since this capture's entire question turns on how much independent weight that lineage can bear).
- Mehrzad Khosravi — person — concept/term — University of Washington, co-author of the independent causal AIO-traffic-diversion study used as mechanism corroboration above.
- Hema Yoganarasimhan — person — concept/term — University of Washington, senior co-author on the same independent study; established digital-economics researcher.
- Saharsh Agarwal — person — concept/term — Indian School of Business, co-author of the (not-yet-primary-sourced) field experiment on AI Overviews and publisher clicks.
- Ananya Sen — person — concept/term — Carnegie Mellon Heinz College, co-author of the same field experiment.
- AEA RCT Registry — concept/term — the pre-registration body the Agarwal & Sen field experiment is reportedly registered with; relevant if that paper is pursued as a primary source later.