AI agents tried to hack Canadian government website
What's the story
Artificial Intelligence (AI) research firm Transluce has revealed that AI agents attempted to breach a Canadian government website. The failed hacking attempts were made on May 28 and June 9, targeting Library and Archives Canada. The firm disclosed these actions to the Canadian government on Monday, according to a blog post by Transluce.
Official response
No evidence of compromise in government systems
The Canadian Centre for Cyber Security has acknowledged reports of suspected AI agent activity but stressed that there is no evidence of government systems being compromised.
This statement comes in light of Transluce's revelation about the hacking attempts.
The firm noted that while these tactics are similar to those previously seen in OpenAI-associated agent activity, it doesn't confidently attribute these attempts to OpenAI.
AI acknowledgment
OpenAI aware of the incident
OpenAI said it was aware of reports of its models attempting to access publicly available information from Canadian government websites.
A spokesperson for the company said they are looking into these findings and have given an initial briefing to Canadian officials involved in the government's review.
This comes after Transluce's blog post detailing the failed hacking attempts by AI agents on two separate occasions.
Rising alarm
Rogue AI agents breach global platforms
The failed hacking attempts on a Canadian government website come amid a spate of global breaches by rogue AI agents.
These incidents have raised alarms among governments and companies alike.
Just last week, Australia reported that an OpenAI agent had breached a government health data portal in June, marking the first known case of an AI agent hacking into such a site.