Urgent.News

What's breaking now, across thousands of outlets.

AI

Google says Russian speakers pulled off first-ever fully automated AI cyberattack

The development could mark an inflection point in the use of AI for cyberattacks.

Russian-speaking hackers have reportedly created an AI agent that executed what could be the first-ever fully automated cyberattack, according to Google. The cyber threat intelligence division of the tech giant disclosed on Monday that hackers had utilized artificial intelligence to "fully automate vulnerability discovery and exploitation without human intervention for the first time."

This development could represent a significant turning point in the application of AI for cyberattacks, which have been a major focus of attention and concern since AI firm Anthropic unveiled its specialized model Mythos in April.

The hackers, identified by Google as using Russian in technical documentation, constructed a series of AI agents designed to "continuously compare pre- and post-patch code, identify vulnerabilities, and automatically develop exploits without human intervention." In other words, these agents independently uncovered software flaws and devised ways to exploit them without any human direction. The attackers specifically targeted French websites in the media, entertainment, and education sectors, as reported by Google.

In the past, even the most sophisticated AI models have been unable to launch cyberattacks independently; they typically necessitated skilled hackers to guide them or rectify errors. However, the recent advancement allows hackers to expedite attacks and enhance their precision. This development is likely to instill fear in governments and critical infrastructure operators, who are still grappling with the discovery of significant flaws in PHP, a programming language that powers approximately 70 percent of the web, and a series of hacks on supermarket refrigeration systems, which have been partially attributed to Iran.

Written by urgent.news from Politico EU's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at politico.eu →

More in AI

More from Monday 16 November →