vulnerability
Google: AI agent turned a security patch into a working exploit with no human help
Google researchers discovered that an AI agent was able to transform a security patch into a working exploit without any human intervention.
Medium