Online scammers use AI programs to generate live deepfake videos, with which they can pretend to be victims' friends during video calls, as found by investigative reporters at China Media Group (CMG).

CMG revealed such scams at the "315" consumer rights gala that aired on Friday.

A victim told CMG that she got a video call from her daughter saying she was kidnapped. But a police investigation found her daughter was fine, and the video call was faked by an AI program.

Police said people can ask the remote person to do a nose touch or shake their head to see if it's generated by AI, as such videos are much harder to generate with AI compared to normal speech with no body language.

What's more, most AI-generated videos cannot last longer than a few seconds with current technology, the police said.

China is one of the first countries to curb "face theft" and regulate the use of generative AI. The government has been requiring AI contents to be labeled since January 2023, which is more than a year earlier than the EU's AI rules.

Watch: CGTN's Guo Meiping and her "twin sister" talk about China's AI rules.