OpenAI launches AI tool to detect security flaws

OpenAI has introduced a new AI-powered cybersecurity platform called “Daybreak,” designed to detect software vulnerabilities, analyze cyber threats and provide automated security solutions.

The company said the platform was developed using its GPT-5.5 model and Codex agent-based coding technology to continuously monitor and strengthen software security systems.

According to OpenAI, Daybreak can identify security flaws, analyze the nature of threats and recommend fixes before systems are compromised, shifting cybersecurity from reactive responses to proactive protection.

Read More: Google launches AI search revolution, ditching old links

The platform is capable of reviewing code, building threat models, verifying security patches, analyzing software dependencies and prioritizing critical vulnerabilities.

Technology analysts say the rapid advancement of AI is transforming cybersecurity, with modern AI models now capable of both enhancing cyberattack capabilities and strengthening defense systems. As a result, competition among major AI companies is expanding beyond chatbots and coding tools into cybersecurity technologies.

Meanwhile, rival AI company Anthropic recently announced its own cybersecurity initiative, “Project Glasswing,” powered by an advanced AI model called Claude Mythos, though its use remains limited for now.