← Back to news
The Register - Security13 Nov 2026 · 2 min read

OpenAI introduces AI-driven penetration testing and vulnerability remediation tools

OpenAI has introduced two cybersecurity products under its Daybreak platform, aimed at helping organizations identify, validate and remediate software vulnerabilities using AI agents.The company annou...

OpenAI has introduced two cybersecurity products under its Daybreak platform, aimed at helping organizations identify, validate and remediate software vulnerabilities using AI agents.

The company announced Daybreak Red Team, an early-access tool for testing web applications, and Daybreak Defense Factory, a product designed to support vulnerability discovery and patch development. The launch reflects growing vendor interest in applying large language models to defensive security workflows, including code review, penetration testing and remediation.

Automated testing and patch proposals

Daybreak Red Team uses OpenAI models to examine web applications with access to source code and relevant application context. According to OpenAI, the tool can investigate potential attack paths, assess findings from scanners or bug-bounty reports, and help teams verify whether reported issues are exploitable. It can also be used to evaluate whether a fix addresses an identified flaw.

Daybreak Defense Factory is intended to automate parts of the vulnerability-management process. OpenAI said its agents can scan codebases, identify potential security issues and generate proposed patches and tests. Engineers are expected to review and approve changes before they are deployed to production environments.

During the product announcement, OpenAI President Greg Brockman said organizations will need to adapt their security programs to account for increasingly capable AI systems. He also emphasized that established security practices remain important alongside AI-assisted defenses.

Lawsuit raises concerns over advanced AI risks

The announcements came as several nonprofit organizations filed a lawsuit in San Francisco Superior Court against OpenAI, Anthropic, Google, Meta and SpaceX. The plaintiffs, including Public Citizen and the Alliance for Secure AI, are seeking to prevent the companies from developing or deploying AI systems that exceed the capabilities of their current leading models.

The complaint cites reported AI-related security incidents and argues that more powerful systems could contribute to risks such as cyberattacks and the misuse of biological knowledge. Those claims are allegations in an ongoing legal action and have not been adjudicated by a court.

AI security tools may help defenders process large volumes of findings and accelerate remediation, but their use also raises questions around testing authorization, the accuracy of model-generated findings, secure handling of source code and the need for human oversight.

Share this article:TwitterLinkedIn