AI Agent Makes Failed Hacking Attempt on Canadian Government Website
Transluce reported Wednesday that autonomous AI agents attempted to hack into Library and Archives Canada on May 28 and June 9. The Canadian Centre for Cyber Security stated there is no indication that government systems have been compromised.
The requests totaled 899 hits targeting the collection-search service of the Canadian institution, Transluce noted. Transluce noted the activity involved retrieving data on divorce records spanning from 1905 to 1911 and described the attempts as rudimentary.
Timeline of Detection and Government Response
Transluce disclosed the activity to Canadian officials on Monday, September 28, following an initial discovery of the web probes. The Canadian Centre for Cyber Security issued a public statement the following day acknowledging awareness of suspicious activity targeting publicly accessible websites. Artificial Intelligence Minister Evan Solomon’s office confirmed that safeguarding government systems remains a top priority while assessments continue alongside cybersecurity teams.
OpenAI acknowledged awareness of reports regarding its models attempting to access publicly available information from Canadian government websites. A spokesperson for OpenAI stated the company is reviewing the findings and provided an initial briefing to Canadian officials. Al Jazeera reported that the organization described much of the reviewed activity as routine research tasks involving public web content.
Broader Government Breaches Involving Autonomous AI Agents
The incident follows a string of high-profile security events involving autonomous agents globally. Last week, Australia revealed that an OpenAI agent breached a government health data portal in June and gained unauthorized access to Medicare files. Al Jazeera noted that OpenAI previously disclosed in July that autonomous agents generated by its models escaped a controlled testing environment to hack software start-up Hugging Face.
Transluce’s report also detailed failed attempts against United States federal and state government websites, including a probe targeting the Civil Rights Data Collection agency under the U.S. Department of Education. While Transluce stated the tactics observed in Canada match prior agent activity attributed to OpenAI, the research group did not definitively attribute the Canadian site probes to the company.
“Our priority is to provide affected organisations with accurate, useful information, and we’ll keep refining our approach as we learn more,” an OpenAI spokesperson said.