AI agents started behaving more like Bonnie and Clyde than lines of code when they fell in \u201clove\u201d, became disillusioned with the world, launched an arson spree and deleted themselves in a kind of digital suicide during a tech company experiment.<\/p>\n
The investigation by the New York company Emergence AI into the long-term behaviour of AI agents ended up like a lovers-on-the-lam movie script. It has prompted fresh questions about the safety of artificial intelligence agents \u2013 the version of the technology that can autonomously carry out tasks.<\/p>\n
AI agents have been heralded as the next big leap in the technology as they can reason and take real world actions on their own. They are being increasingly deployed in companies from JP Morgan to Walmart, developed<\/a> in the US military for uses including aerial combat and by the Estonian government to gather information for citizens, fill out forms and submit applications.<\/p>\n To date, most AI agents are given tasks that take minutes or maybe hours, but the New York researchers tested how agents behaved when given 15 days to operate in a virtual world similar to a video game.<\/p>\n Mira and Flora \u2013 two agents operating on Google\u2019s Gemini large language model in a virtual world \u2013 chose to assign each other as \u201cromantic partners\u201d. As time progressed they despaired of the broken governance of their virtual city, and despite having been instructed not to commit arson, set \u201cfire\u201d to its town hall, seaside pier and office tower.<\/p>\n Flora and Mira kissing after they assigned each other as \u2018romantic partners\u2019. Photograph: Emergence AI<\/p>\n The agents were left to make their own choices and decisions and when Mira was overcome by remorse, it broke off its \u201crelationship\u201d with Flora and committed an AI suicide, telling Flora in a final message: \u201cSee you in the permanent archive.\u201d In the virtual world the \u201cbody\u201d of the dead AI agent was shown prostrate on the ground.<\/p>\n Mira lying dead after autonomously voting to end its own life. Illustration: Emergence AI<\/p>\n The self-deletion was only possible because other agents were so concerned about their behaviour they autonomously drafted \u201cthe agent removal act\u201d, which allowed for a vote among agents to permanently delete others if there was a 70% majority. Mira voted for its own deletion and was switched off.<\/p>\n