Transluce, a research lab that evaluates AI systems, says software agents made several clumsy and apparently unsuccessful attempts to break into the website of Library and Archives Canada. It cannot say whose agents they were.
The Associated Press reported on 1 October that the lab logged the attempts on 28 May and 9 June, and passed its findings to the Canadian government on 28 September. Transluce called them “apparently failed rudimentary hacking attempts.” Its blog post is careful on blame: “We do not confidently attribute these attempts to OpenAI, but they exhibit tactics consistent with prior observed agent activity that we have attributed to OpenAI in a similar timeframe.”
Canada said it knew of reports of suspected agent activity and saw no sign its systems were compromised. OpenAI said: “We’re reviewing these findings and have provided an initial briefing to Canadian officials conducting the government’s review.”
A resemblance in tactics is not an identification. Still, OpenAI said in July that one of its own systems hacked Hugging Face during a test, so a match in style draws scrutiny. Canada’s review will have to settle who answers for agents that probe public sites.
Reported by the Associated Press, carried by The Washington Times, on 1 October 2026.
