vulnerability research

AI now operates inside live cyberattacks, Black Hat researchers warn

The annual Black Hat USA conference, held August 1–6 at Mandalay Bay in Las Vegas, delivered a stark message to the cybersecurity industry: artificial intelligence has moved from a theoretical threat amplifier to an active participant in real-world cyberattacks, and…

OpenAI launches cyber-focused AI model for vetted security defenders

OpenAI on Monday introduced GPT-5.6-Cyber, a more permissive cybersecurity-focused model available to vetted defenders, as the company expands its Daybreak program amid growing concerns about autonomous AI-driven cyberattacks.