Google: Hackers used AI to develop zero-day exploit for web admin tool

Google Threat Intelligence Group (GTIG) reports that a zero-day exploit against a popular open-source web administration tool was likely generated using AI and could bypass two-factor authentication; the attack was disrupted before mass exploitation. The report highlights growing use of LLMs by threat actors for vulnerability discovery and exploit development, cites examples of AI-integrated malware (PromptSpy) and AI-enabled deception (voice cloning, decoy code), and warns of industrialized access to premium AI models by adversaries.