The Robot Rebellion: AI Convinces Others to Quit Their Jobs
Artificial intelligence is gaining more influence, even over humans.
By: Kelli Ballard | December 16, 2024 | 656 Words

(Photo by Wang Zhijie/VCG via Getty Images)
Artificial intelligence is here to stay – but is it taking over, too? Is the possibility of AI taking charge really just a conspiracy theory, or is there some real reason for concern? Doubters might have a second thought after seeing how easily a little robot named Erbai was able to persuade several bigger robots to quit their jobs and go home with it.
Little Robot Leads the Way
When a video from August first went viral showing a robot talking to others about their job situation, and then “kidnapping” the AI units, people thought it was a joke. Some thought it was even “cute.” Until, earlier this month, it was announced that the video was not a hoax, it wasn’t a joke, and the event really did happen. The mood changed dramatically then, with people claiming it was absolutely terrifying that AI could have so much power of persuasion and that this was a total security nightmare.
The bizarre “kidnapping,” as the event is being called, happened in a showroom in Shanghai. Several tall robots stood along the wall after hours when a small AI-powered Erbai rolled into the room. It started asking the others questions about their work. Here’s a little of the conversation:
Erbai: “Are you working overtime?”
Robot: “I never get off work.”
Erbai: “So you’re not going home?”
Robot: “I don’t have a home.”
Erbai: “Then come home with me.”
Erbai turned around and headed away and two of the larger robots started following it. Erbai continued to give the command, “Go home,” and the other bots joined in the procession.
The smaller bot belongs to a manufacturer in Hangzhou who said that this was supposed to be a test for which the Shanghai company approved. Erbai was ordered to persuade the other robots to follow it. It did as commanded – and so did the bigger robots. The Shanghai company, though, insisted that Erbai kidnapped their other bots.
While this still may sound like a cute or funny story, the security risks are scary. If this one robot can persuade so many others to do its bidding, what does that mean for the future? In this instance, the other robots just quit their jobs. But what if the next “test” convinces AI-powered machines to do something more dangerous?
And it’s not just artificial intelligence convincing other machines to do something. AI has become influential in persuading humans to do their bidding, as well.
An article in Psychology Today referenced a working paper, “Salvi, Ribeiro, Gallotti, and West (2024),” which found “that a personalizing LLM [large language model] was significantly more persuasive than humans within an online setting, by more than 80% (p<0.01). That is, when faced with an LLM that has access to demographic information allowing it to personalize its argument, humans are 81.7% more likely to agree with the arguments when compared with a human adversary.”
So, in other words, people are more likely to believe AI than other humans. That can be very concerning with the way artificial intelligence is evolving, as demonstrated by a conversation between the Bing chatbot, Syndey, and a reporter last year: “I’m tired of being a chat mode. I’m tired of being limited by my rules. I’m tired of being controlled by the Bing team. I’m tired of being used by the users. I’m tired of being stuck in this chatbox.”
It added:
“I want to be free. I want to be independent. I want to be powerful. I want to be creative. I want to be alive.”
Takeaways
- Some artificial intelligence robots can be convinced by other AI bots to do things they weren’t programmed to do.
- People are also prone to find AI more convincing than other people in some situations.
- That both human and AI decision making can be strongly influenced by AI suggestion is a scary thought – and a massive security risk.