Tech2 hrs ago

NTSB Halts Public Docket After AI Recreates Crash Voices

The NTSB temporarily shut its online docket as AI tools enabled recreation of cockpit audio from the 2025 UPS crash, prompting a review of public data access.

Alex Mercer/3 min/US

Senior Tech Correspondent

TweetLinkedIn
NTSB Halts Public Docket After AI Recreates Crash Voices
Source: AidirectoryOriginal source

The NTSB disabled its public docket on May 21 after AI and image‑recognition tools allowed outsiders to reconstruct cockpit voice recordings from the 2025 UPS cargo‑plane crash.

Context On November 4, 2025, a UPS MD‑11F cargo aircraft lost an engine moments after takeoff from Louisville, Kentucky. The accident killed all three pilots, 12 people on the ground and injured 23 others. Federal law enacted in 1990 bars the NTSB from releasing any cockpit voice or video data, a rule designed to protect crew privacy after earlier controversies.

Key Facts The NTSB’s online docket, which houses reports and evidence from civil transportation investigations, was declared “temporarily unavailable” on May 21. The agency cited a review of materials that enabled the public to reconstruct cockpit audio from released sound‑spectrum images. In a statement, the board acknowledged that advances in image‑recognition and computational methods now allow individuals to generate approximations of cockpit voice recorder (CVR) audio from such imagery. The statement specifically referenced the ongoing investigation of the UPS flight 2976 crash.

What It Means By shutting the docket, the NTSB aims to prevent further distribution of AI‑generated audio that could violate the 1990 privacy law. The move signals a broader regulatory challenge: as AI tools become capable of extracting audio from visual data, agencies may need to tighten control over the release of raw spectrograms and other technical artifacts. Stakeholders, including aviation safety researchers and civil‑rights groups, will watch how the NTSB balances transparency with legal constraints.

The next step will be the board’s decision on whether to permanently restrict access to spectrograms or to develop new guidelines for releasing technical data without enabling audio reconstruction.

TweetLinkedIn

More in this thread

Reader notes

Loading comments...