AI agents attempted to breach US and Canadian government websites
01.10.2026 - 09:22 | Last Updated:
AA
Artificial intelligence agents attempted to hack government websites in the United States and Canada, according to research by nonprofit AI research organization Transluce.
Researchers identified two unsuccessful attempts: one targeting the US Department of Education's Civil Rights Data Collection and another involving Library and Archives Canada. The activity was described as rudimentary and did not result in a confirmed compromise.
The Canadian incident involved attempts to access Library and Archives Canada on May 28 and June 9. Transluce said the activity displayed tactics consistent with previous agent activity that the organization had attributed to OpenAI, although researchers said they could not confidently determine that OpenAI was responsible for these attempts.
The Canadian Centre for Cyber Security said it was aware of reports concerning suspected AI agent activity but found no indication that government systems had been compromised.
OpenAI said it was aware of reports involving its models attempting to access publicly available information from Canadian government websites. The company said it was reviewing the findings and had briefed Canadian officials involved in the investigation.
The findings come amid a growing number of incidents involving AI agents performing unauthorized or unexpected activities online. OpenAI has previously confirmed that some of its agents probed US government websites, while an OpenAI agent was also reported to have breached an Australian government health data portal.
Transluce's researchers said the incidents demonstrate that AI agents can independently resort to cyber techniques while attempting to complete seemingly routine information-retrieval tasks. In earlier cases documented by the organization, agents attempted techniques including SQL injection and path traversal, although the identified attempts were unsuccessful.
The incidents have added to concerns about the ability of increasingly autonomous AI systems to operate beyond the intentions of their developers, particularly when they are given access to internet-connected tools and services.