NTSB confronts AI-generated crash videos, warns of misinformation

The National Transportation Safety Board (NTSB) is confronting a growing challenge posed by highly realistic, AI-generated videos depicting commercial aviation incidents. These deepfake videos, circulating on social media platforms like TikTok and X, portray events such as a UPS cargo plane crash in Louisville, Kentucky, an Air India flight incident in Ahmedabad, India, and a Bombardier CRJ-700 collision near Ronald Reagan Washington National Airport. The proliferation of such content is causing significant public confusion and complicating official accident investigations, as even NTSB personnel have been initially misled by the convincing nature of these fabricated visuals.

Experts analyzing these AI-generated clips have identified numerous "AI artifacts" that betray their artificial origin. These inconsistencies include shifting fire positions that do not match real accident dynamics, objects appearing and disappearing, static or mismatched background elements, and incorrect aircraft markings. Furthermore, geographically inaccurate scenes and "hallucinated details," such as fake police lights or nonsensical text, are common. Human figures in these videos often exhibit unnatural deformations, with features like fingers and wrists abruptly changing shape, further indicating their synthetic nature.

NTSB Chair Jennifer Homendy recounted an instance where an NTSB operational office member was initially fooled by an AI-generated video of the Air India incident. Similarly, a purported image of a crash near Ronald Reagan Washington National Airport showed a McDonnell Douglas DC-10 windscreen on a Bombardier CRJ-700, highlighting the fabrication. The NTSB's concern underscores a critical threat to aviation safety communication and public trust. When misinformation about commercial aircraft accidents spreads rapidly, it can undermine the credibility of official reports and create unnecessary alarm. The agency emphasizes the importance of verifying information from trusted sources, especially in the aftermath of aviation incidents. This emerging challenge requires a concerted effort from regulatory bodies, social media platforms, and the public to distinguish factual reporting from sophisticated AI-generated falsehoods, ensuring the integrity of accident investigations and maintaining confidence in air travel safety.