Google Says AI Agents Have Run Their First Fully Autonomous Cyberattacks
Google’s threat researchers say they have watched AI frameworks find and exploit software flaws end to end with no human in the loop — a threshold they warn will shrink the time defenders have to patch.
