Artificial intelligence (AI) has transformed cybersecurity from a mostly human-led process into an increasingly automated battlefield. Today, companies are deploying AI systems to detect threats, analyse attacks, and automatically fix vulnerabilities in real time. Against this backdrop, OpenAI on Monday unveiled Daybreak, its latest cybersecurity initiative and strongest push yet into AI-powered cyber defence.

OpenAI’s latest announcement comes a month after Anthropic announced Project Glasswing, its cybersecurity initiative powered by an unreleased frontier AI model named Claude Mythos. With advanced AI models becoming powerful enough to both create and defend against cyber threats, top AI companies are pivoting towards cybersecurity. This shift is happening at a time when frontier models are showing dramatic improvements in coding, reasoning, and autonomous task execution.