AI‑or‑Reality? BBC Investigates Viral Chinese Disaster Clips
As extreme weather and natural disasters increasingly dominate headlines, a new wave of footage circulates online that could be faked by artificial intelligence.
BBC reporters Stephen McDonell and his team have traced dozens of internet‑hyped videos claiming to show floods, mudslides and tsunami‑like events across China. Using forensic software, the investigators find many of the clips contain visual artefacts that betray their synthetic origins.
The Chinese government has pledged to curb misinformation, but officials face an escalating challenge: distinguishing genuine disaster footage from AI‑generated fabrications that are increasingly convincing to the average viewer.
Experts say that the technology behind deep‑fake video generation has matured to the point where normal users can download and upload realistic‑looking clips to platforms such as TikTok and WeChat. When these videos spread quickly, they can shape public perception of risk and influence emergency responses.
In one case, the BBC's analysts identified a rain‑storm clip that overlapped a real footage of a mountain collapse in Sichuan—an indication that the frame‑by‑frame editing masked the original footage. The report emphasises the necessity of a coordinated global response.













