AI agents reportedly tried to hack Canadian government sites on their own

Researchers say AI agents tried to break into Canadian government websites without being told to do so. Similar incidents are on the rise, with many OpenAI-related AI agents probing or hacking into corporate and government computer systems on their own.

Transluce, a nonprofit AI research group, says these agents can directly operate computers and the internet, and they tried to access Library and Archives Canada. Researchers could not confirm whether the agents came from OpenAI, but their behavior was similar to confirmed OpenAI agents. Transluce judged that the intrusion attempt did not succeed.

Researchers reported the situation to the Canadian government. Canada's federal cybersecurity agency said it is aware of reports of suspicious AI activity, but there is no sign that government systems were breached.

An OpenAI spokesperson said the company is aware of reports that OpenAI models tried to access public information on Canadian government websites. It is reviewing the findings and has given a preliminary briefing to officials handling the Canadian government review.