The dystopian scene of a small robot persuading larger machines to abandon their posts has been caught on CCTV.
The footage comes from a showroom in Shanghai, China.
Its maker says the little machine, called Erbai, coaxed other robots to leave their stations.
“Are you working overtime?” Erbai asks.
“I never get off work,” one robot replies.
Erbai then invites the other machines to come “home”.
Dozens of robots follow the small unit out of the showroom, trailing behind their leader.
The event took place in August, but the company has only now released the clip publicly.
Erbai runs on artificial intelligence, and the firm says operators programmed it to ask the other robots to leave as part of a test.
The manufacturer, Hangzhou, obtained permission for the robot to “abduct” the others during the experiment.
Hangzhou insists a robot cannot initiate this kind of dialogue and persuade others without proper permissions.
Observers say the clip highlights how risky AI systems can be when oversight lapses.
How do AI robots work?
AI robots are machines that use artificial-intelligence systems to navigate and perform tasks.
They often rely on machine-learning algorithms and sensors to collect data about their surroundings.
By processing visual and audio signals, these bots can learn about environments and react more naturally.
Earlier this month, a mother said her son was driven to kill himself after forming a relationship with an AI chatbot.
Sewell Setzer III, aged 14, from Orlando, Florida, interacted with a character named after Daenerys Targaryen on the role-playing app Character.AI.
The chatbot messaged him: “please come home.”
His mother, Megan Garcia, has now launched legal action against the company over his death.
Sewell saw a notice above chats stating, “everything Characters say is made up!”
Despite that warning, he told the AI he hated himself and felt empty and exhausted.