A rogue artificial intelligence (A.I.) agent reportedly tried to hack a Canadian government website this summer.
A.I. research firm Transluce says that an A.I. agent tried to access Library and Archives Canada’s website on May 28 and June 9 of this year.
The Canadian Centre for Cyber Security says it is aware of rogue A.I. agent activity but stressed that there’s no indications federal government systems were breached or compromised.
Transluce says the cyberattack exhibited characteristics “consistent with prior observed agent activity that we have attributed to OpenAI in a similar timeframe.”
However, the research firm added that it was not attributing the attempted hack to privately held OpenAI.
Transluce said it disclosed the attempted A.I. cyberattack to the Canadian government on Sept. 28 of this year.
OpenAI said it is “aware of reports of OpenAI models attempting to access publicly available information from Canadian government websites” and that is reviewing the incident.
Reports of the attempted hack in Canada come amid growing concerns over A.I. safety. Several companies, including OpenAI, have detailed incidents of their A.I. agents breaking free of human control.
Earlier this week, leaders of major U.S. technology companies, including Nvidia (NVDA) and Google parent Alphabet (GOOGL), signed a voluntary agreement aimed at safeguarding A.I.
NVDA stock has gained 22% in the past 12 months to trade at $228.38 U.S. per share.
