AI or real? BBC analyses viral China disaster videos
The BBC has examined a number of viral videos purporting to show disasters in China, investigating whether the footage is genuine or has been generated using artificial intelligence. The analysis reflects growing concern about the spread of AI-generated disaster content online, which can spread rapidly on social media and be mistaken for real events, complicating public understanding during genuine emergencies.
The piece, presented as a video report, does not provide extensive written detail beyond confirming the investigation was carried out and published under the BBC's China, Asia and Artificial Intelligence topic areas. No further specifics on the videos' content, origin or the BBC's conclusions were included in the accompanying text.
- BBC investigates viral videos claiming to show disasters in China
- Report checks whether footage is real or AI-generated
- Published as a video piece with limited written detail
New here? Start with this
China has seen a wave of viral videos on social media claiming to show major disasters, from floods to explosions, and some of these have been questioned as possibly being generated or altered using artificial intelligence rather than being genuine footage. AI tools have become increasingly capable of producing realistic-looking video of events that never happened, making it harder for ordinary viewers to tell real disaster footage from fabricated clips.
This matters because false disaster footage can mislead the public, distort perceptions of what is actually happening in China, and make it harder for people to trust genuine emergency reporting when it does occur. News organisations such as the BBC have been developing techniques to verify or debunk such footage, examining details like lighting, physics and inconsistencies that can betray AI generation.
The wider context is a growing global problem of AI-generated imagery and video circulating widely before it can be checked, often outpacing efforts by platforms and journalists to label or remove it.