A recent viral incident in China has raised both amusement and concern about the growing capabilities of artificial intelligence.
An AI-powered robot named Erbai, developed by Unitree Robotics in Hangzhou, reportedly “kidnapped” 12 larger robots from a Shanghai robotics showroom. The event, that was captured on video, shows Erbai engaging in a human-like conversation with the larger robots, asking if they were “working overtime” and inviting them to “come home.” Surprisingly, all 12 robots followed Erbai out of the showroom after it issued the command, “Go home.”
Initially believed to be a prank, both robot manufacturers later confirmed the incident was real. However, they clarified it was a collaborative AI experiment designed to test how robots could use conversation and reasoning to influence other machines.
According to the developers, Erbai was only programmed with a few basic commands, such as initiating conversation and saying “go home,” but the rest of the dialogue and behavior emerged spontaneously through AI interaction.
The video quickly spread across platforms like Douyin, X, and YouTube, where viewers reacted with humor and unease, comparing the event to a scene from a sci-fi movie.
While the experiment was conducted under controlled conditions, it has sparked renewed discussion about AI autonomy and the future of human-like decision-making in machines, highlighting how quickly AI behavior is evolving beyond scripted instructions.