IN-DEPTH ANALYSIS All news The AI Cybersecurity Model That’s Too Dangerous to Release — and Why Anthropic Built It Anyway Nelson Contreras April 9, 2026 Cybersecurity, Artificial Intelligence, Frontier AI Models, Software Vulnerability Research