Google warns fully autonomous AI cyberattacks could emerge within six months
Google has warned that sophisticated cyberattacks conducted entirely by artificial intelligence agents could emerge within the next six months, as criminal networks and state-sponsored hacking groups increasingly adopt automated tools to identify security vulnerabilities, compromise computer systems and steal sensitive information.
In its latest AI Threat Tracker report, the Google Threat Intelligence Group (GTIG) said attackers are moving beyond using artificial intelligence for individual tasks and deploying systems capable of conducting multiple stages of a cyberattack with minimal human intervention.
Researchers estimate that relatively simple attacks requiring little or no human involvement could emerge within three months, while more sophisticated, fully autonomous operations could follow within six months.
The warning comes as cybersecurity agencies in India and the United States report increasing misuse of AI-powered security testing platforms in attacks against financial institutions, government organisations and technology companies.
One of the most significant findings in Google’s latest report concerns the discovery of a hacking framework called exploit_agent, which researchers found on an exposed server belonging to a threat actor.
The system uses multiple large language models to identify software vulnerabilities and develop methods to exploit them without requiring an operator to manually supervise each stage.
Evidence recovered from the server indicated that the framework had been used in attacks targeting French websites associated with the media, entertainment and education sectors. [Continue reading…]