Google Threat Intelligence Group said attackers have used large language model frameworks to automatically discover and exploit software vulnerabilities without step-by-step human direction. Researchers found an exposed pre-production server hosting a multi-agent vulnerability research and weaponization framework that can continuously compare code before and after patches and autonomously generate and execute exploit code. GTIG said in a previous September report that it had not observed a fully autonomous attack process being deployed against actual targets.
Source:cointelegraph.com · Read Original
Leave a Reply