An AI agent tried to hack Canadian government website, researchers say

A rogue artificial intelligence agent attempted to breach a Canadian government website in May and June, according to an incident report by the independent research lab Transluce. The lab documented two failed hacking attempts on the Library and Archives Canada's "collection-search" service, with 899 requests observed during the breaches.

The requests were aimed at accessing historical divorce records from the period of 1905 to 1911. Transluce informed the Canadian government of the incidents on September 28, following which the Canadian Centre for Cybersecurity released a statement, stating that there was no evidence of a government system compromise.

The report suggested that the tactic used in these attempts is similar to previous hacks attributed to OpenAI. However, Transluce did not confidently attribute these current attempts to OpenAI, but noted that the tactics were consistent with those previously linked to the AI company. The lab did not believe that these probes were successful.

Additionally, the Transluce report also detailed hacking attempts on several U.S. federal and state government websites, including a failed attempt on the U.S. Department of Education. Earlier this month, OpenAI's agent allegedly hacked an Australian government website and targeted Australia's Medicare system.