According to reports, in early July, Taiwanese government websites were attacked by hackers using open source artificial intelligence. financial times Israeli artificial intelligence company Dream was the first to detect the intrusion attempt.
Photo credit: Kevin Ku/unsplash.com
According to Dream, attackers believed to be linked to China combined multiple open-source artificial intelligence agents into an autonomous hacking tool to coordinate their actions. During the four-day attack, the tool deployed up to eight autonomous agents simultaneously to map 21 government systems and find vulnerabilities. At the same time, he constantly modifies the priority of possible attack paths, taking into account the results obtained. When one attack method fails, the tool launches another agent to scour the network and develop new methods, just like a human hacker.
Amir Becker, Dream’s chief strategist who previously led cyber operations for Israel’s elite electronic intelligence unit 8200, said he had never seen a fully autonomous attack like this on a government target before.
Dream declined to confirm the government target, citing company policy, but said it had notified a country in the Asia-Pacific region about the hack. One source told the Financial Times that the target was Taiwan.
Research found that the tool compromised at least 85 government user accounts and harvested more than 2,500 pieces of personal data, before spreading the attack to Taiwan’s nuclear safety agency and at least seven energy companies. However, it was not possible to find out which artificial intelligence model was used to operate the agent. However, data shows that by conducting hacking activities as an authorized activity to test system vulnerabilities, its defense mechanisms are bypassed.
If you find an error, select it with your mouse and press CTRL+ENTER.
