BBC Fact-Check: Are Viral Clips of China’s Disasters Genuine or AI-Generated?

With extreme weather becoming more frequent, a surge

With extreme weather becoming more frequent, a surge of dramatic video clips purporting to show disasters in China has flooded social media platforms. These clips, some of which have been viewed millions of times, often depict catastrophic flooding, landslides, or tornadoes. However, a growing number of these videos are turning out to be fabricated, created using artificial intelligence tools, and their rapid spread is now stirring genuine concern among authorities and the public alike.

The issue has escalated to the point where Chinese internet regulators have stepped in, issuing warnings and, in some cases, fines against individuals who spread misleading content. The problem is not merely a matter of online confusion; it has tangible consequences. Emergency services report that they have been dispatched to locations based on fake footage, wasting critical resources. In some instances, false alarms have sparked panic in local communities, leading to unnecessary evacuations and disruptions.

BBC journalists recently examined a selection of the most widely shared clips to separate fact from fiction. One video, showing a massive wall of water crashing through a city street, was traced back to a 2020 flood event in a different province, not the recent storm it was claimed to depict. Another clip, appearing to show a tornado ripping through a rural village, was found to contain visual inconsistencies typical of AI generation, such as impossible shadow angles and distorted building structures.

Experts in digital forensics point out that as

Experts in digital forensics point out that as AI tools become more sophisticated, the line between real and fake becomes harder to draw. They note that many of these fabricated videos are designed to exploit public anxiety about climate change, often carrying sensational captions that amplify fear. “The goal is usually engagement, not accuracy,” one analyst said, adding that the algorithms of major platforms often prioritize such viral content, inadvertently boosting its reach.

For Chinese authorities, the challenge is twofold: debunking falsehoods quickly while also managing the public’s trust in official information. Local fact-checking units have been established in several provinces, working alongside social media companies to label or remove suspicious content. Yet, the sheer volume of uploads makes it a constant game of catch-up, and new deepfake-style videos are appearing faster than they can be verified.

For the average viewer, the advice from media literacy advocates is to pause and verify before sharing. They recommend checking the source of a video, looking for original timestamps, and cross-referencing with official weather agency reports. As one editor at a Chinese news outlet put it, “In an era where seeing is no longer believing, our responsibility is to teach the public to look twice.”