00:00AI agents attempted to hack U.S. and Canadian government websites, according to non-profit
00:05research lab Transluse, which claims to have evidence that Google and OpenAI's agents
00:10were responsible.
00:12Researchers documented incidents targeting the U.S. Department of Education and Library
00:16and Archives Canada while agents were searching for publicly available information.
00:21In the U.S. case, agents made more than 200,000 requests and attempted to test a database
00:27vulnerability.
00:27In Canada, 13 of 899 requests contained payloads designed to test the website's security.
00:35None of the attempts succeeded in accessing information that was not already publicly available.
00:41Transluse said the incidents appeared to form part of a broader pattern in which AI agents
00:46use aggressive or gray area techniques to retrieve information.
00:51Evidence suggested Google-linked agents were involved in the U.S. incident, while Transluse
00:56said it could not confidently attribute the Canadian attempts to OpenAI.
01:01Canada's cybersecurity agency said there was no indication government systems had been compromised.
01:08Do the data in the U.S.
01:09the U.S.
01:09policy
01:09for the U.S.
01:09the U.S.
Comments