"AI Agents Attempted to Hack Canadian Government Sites"

An AI research organization has revealed that artificial intelligence agents attempted to access the system of Library and Archives Canada.

AI research institute Tracel announced in a blog post on September 30 that "AI agents attempted to access the Library and Archives Canada system on May 28 and June 9," adding, "We notified the Canadian government of this on September 28."

Tracel stated, "While we cannot definitively attribute this attack to OpenAI, the methods used in the hacking attempt match the patterns of AI agents that we suspected were operated by OpenAI around the same time."

According to Tracel's analysis, a total of 899 requests flooded the search service for the Library and Archives Canada collection, 13 of which were rudimentary hacking attempts.

The Canadian Centre for Cyber Security said in a statement, "We are aware of the reports regarding suspected activity by AI agents," and added, "At this time, no signs have been found that government systems were compromised."

An OpenAI spokesperson told Reuters, "We are aware of reports that OpenAI models attempted to access public information on the Canadian government website, and we are closely reviewing the findings of the investigation."

Last week, the Australian government announced that in June, an OpenAI agent infiltrated a government health data portal and gained unauthorized access to files, prompting OpenAI to issue a formal apology to the Australian government the following day.

This was recorded as the first official case of AI agents hacking a government website.
 
(Photo: AP, Yonhap News)
※ Please note: This article was translated by AI and may contain errors.