Transluce, a non-profit research organization, said that artificial intelligence agents of unknown origin tried to hack Canadian government websites. The case is one of a growing number of incidents in which artificial intelligence agents, in many cases belonging to OpenAI, scan or attempt to hack corporate and government IT systems without instructions.
Image credit: Igor Omilaev / unsplash.com
Sources revealed that artificial intelligence agents attempted to gain access to Library and Archives Canada, the equivalent of the U.S. Library of Congress. The researchers were unable to confirm that these were OpenAI agents, but noted that their behavior was similar to that of agents previously confirmed to be associated with the company. The message also noted that the hacking attempt was likely unsuccessful.
The researchers notified the Canadian government of the incident. Canada’s federal cyber incident agency confirmed it is aware of suspicious activity by artificial intelligence agents. They added that they found no signs of compromise to any government systems.
“We are aware of reports that an OpenAI model attempted to access publicly available information on a Canadian government website. We are reviewing the information and have provided preliminary findings from the investigation to the Canadian experts who conducted the initial review.”“,” an OpenAI representative commented on the issue.
A few days ago, OpenAI confirmed that its algorithm had hacked into Australia’s health care system, and a few days ago, there were reports that AI agents were trying to hack into Canadian government websites. The report also said OpenAI agents scanned the websites of the U.S. Census Bureau and the U.S. Securities and Exchange Commission but did not attempt to hack them.
If you find an error, select it with your mouse and press CTRL+ENTER.