A growing number of women in China are reportedly using artificial intelligence (AI) to generate videos and images of homeless men inside their homes to test the depth of their partners’ affection, according to a report by South China Morning Post.

(Also read: Chinese man drops $14,000 gold straw on way home, police find it: ‘Now my wife will not punish me’)

These women, active on mainland Chinese social media platforms, said they asked AI tools to create realistic footage of vagabonds entering their rooms. They then sent the fabricated videos to their husbands or boyfriends to see how they would react. Some women even crafted stories claiming they had invited a homeless man into their home for food and shelter.

Several of these AI-generated clips went viral online, with some posts receiving more than 400,000 likes, SCMP reported.

Husband calls police after AI Prank

One incident on 17 October in eastern China’s Anhui province escalated beyond expectation. A woman sent AI-generated images of a homeless man inside her home to her husband, who was out dining with friends.

Believing the footage to be real, the man immediately alerted the police. Officers rushed to their home, only to discover that the alarming images were fabricated using AI technology. They admonished the woman for “fear mongering” and “wasting public resources,” SCMP reported.

Legal warnings

The trend has drawn strong criticism online, with some women admitting regret. One woman wrote in an October post that she would “never try AI homeless pranks again” after her partner involved property management staff to check on her wellbeing.

Another user claimed that a similar prank in her neighbourhood led to a police complaint, and the woman responsible was punished for spreading false information.

Under China’s Public Security Administration Punishment Law, individuals engaging in such acts can face up to 10 days of detention and a fine of 500 yuan (US$70), SCMP reported.

Authorities also warned that if such AI-generated footage spreads widely and triggers public panic, it could be classified as fabricating false information, an offence punishable by up to five years in prison.