TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
AI technology was used to recreate the voices of pilots who died in a 2025 UPS plane crash from spectrogram images and transcripts. The NTSB has temporarily restricted access to its investigation docket to review the implications. This development raises concerns about privacy, ethics, and the potential misuse of AI in aviation investigations.
The National Transportation Safety Board (NTSB) has temporarily restricted access to its investigation docket after discovering that AI was used to reconstruct the voices of pilots killed in a 2025 UPS plane crash, raising ethical and privacy concerns.
According to the NTSB, the investigation docket for the UPS Flight 2976 crash included a spectrogram file of the cockpit voice recorder, which is normally protected by federal law from public release. Tech-savvy individuals used AI tools, such as OpenAI’s Codex, alongside publicly available transcripts and the spectrogram, to approximate the original cockpit audio, as noted by YouTuber Scott Manley on X.
The NTSB confirmed that the reconstructed voices were created from data that was publicly accessible, and that this activity was not authorized by the agency. AI is being used to resurrect the voices of dead pilots. The investigation into the crash remains ongoing, with 42 cases, including this one, temporarily closed pending review. The agency has not specified whether the AI-generated voices will be used in official findings or if they influenced the investigation’s progress.
Why It Matters
This incident highlights the growing capabilities of AI to reconstruct sensitive audio from limited data, raising questions about privacy, data security, and the ethical use of AI in aviation and accident investigations. It also prompts a review of policies governing the handling and release of investigation data, especially cockpit recordings.
AI audio reconstruction software
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
The UPS Flight 2976 crash occurred in Louisville, Kentucky, in 2025, resulting in the deaths of both pilots. The investigation had been progressing with access to various flight data, but cockpit voice recordings are typically protected by law from public disclosure. The spectrogram file, a visual representation of sound frequencies, was included in the publicly accessible investigation docket. Tech communities and AI enthusiasts have demonstrated the ability to reconstruct audio from such visual data, as shown by Scott Manley’s analysis. This development underscores the increasing intersection of AI technology with sensitive investigative information. AI is being used to resurrect the voices of dead pilots.
“We are reviewing the use of AI to reconstruct cockpit voices and are considering necessary policy updates to protect investigation integrity and privacy.”
— NTSB spokesperson
“Using publicly available spectrograms and transcripts, AI can approximate cockpit audio, which raises important questions about data security.”
— Scott Manley
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
It is not yet clear whether the AI-generated voices will be officially used in the investigation or if this activity will lead to regulatory changes. AI is being used to resurrect the voices of dead pilots. The extent of the AI’s accuracy and the potential for misuse remain under review.
As an affiliate, we earn on qualifying purchases.
What’s Next
The NTSB is expected to conduct a comprehensive review of investigation data handling policies and may implement stricter controls on sensitive information. Further updates on whether AI reconstructions will influence official findings are anticipated in the coming weeks.
As an affiliate, we earn on qualifying purchases.
Key Questions
Could AI-generated voices impact the investigation’s conclusions?
It is currently unclear if the reconstructed voices will be used in official findings, as the investigation is still ongoing and under review.
Does this mean AI can create realistic voices of anyone from limited data?
Yes, current AI tools can approximate voices from visual data like spectrograms, but the accuracy and ethical implications are still being evaluated.
What are the privacy concerns related to this development?
This raises questions about the privacy of individuals involved in investigations and the potential misuse of publicly available data to recreate sensitive audio.
Will the NTSB change its policies after this incident?
The agency has indicated it is reviewing its data handling policies and may implement new safeguards to prevent unauthorized reconstruction of investigation materials. AI is being used to resurrect the voices of dead pilots.
Source: TechCrunch
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
