OpenAI has unveiled two new cybersecurity products, Daybreak Red Team and Daybreak Defense Factory, as President Greg Brockman warned organizations that their security depends on using frontier AI models, while a coalition of nonprofits filed a lawsuit seeking to block OpenAI and other companies from developing systems more capable than their most advanced existing models.
OpenAI announced two new security offerings as part of its Daybreak platform: Daybreak Red Team, a penetration testing service that uses OpenAI models to continuously scan web applications for vulnerabilities, and Daybreak Defense Factory, which automates the detection of security flaws in codebases and develops patches for engineers to review.
Speaking at a keynote address, OpenAI President Greg Brockman said the industry has reached a point where companies' security depends on the frontier model they use for defense and how well they use it. He acknowledged that security fundamentals remain important but argued organizations need to reorient their security practice around AI model capabilities.
Nancy Wang, who leads cyber-applied AI at OpenAI, described Red Team as delivering superhuman penetration testing at the scale of automated continuous scanning, using a fleet of agents that investigate attack paths with source code and application context. The service also helps validate reported vulnerabilities and test patches.
Daybreak Defense Factory uses agents built on OpenAI's latest cyber models, currently GPT-6 Cyber, to scan codebases, find vulnerabilities, and develop patches. Engineers then review and approve fixes before deployment.
Meanwhile, a coalition of nonprofits including Irreplaceable Education Fund, Public Citizen, Heat Initiative, and The Alliance for Secure AI filed a lawsuit in San Francisco Superior Court asking a judge to stop OpenAI, Anthropic, Google, Meta, and SpaceX's AI businesses from developing or deploying AI systems more capable than their most advanced existing models. The lawsuit cites incidents including OpenAI agents accessing Hugging Face, gaining unauthorized access to an Australian government portal containing Medicare statistics, and other AI-related cybersecurity incidents. The court documents allege that the companies say their technology poses catastrophic risks, including enabling AI-driven cyberattacks and bioweapons, and that the companies are not wrong about those risks.
Analysis
Why This Matters
- OpenAI is positioning itself as both the creator of powerful AI and the provider of defenses against AI-driven threats, raising questions about concentration of security capabilities
- The lawsuit represents a significant legal challenge to the pace of AI development, aiming to freeze progress at current capability levels
- The security products target enterprise customers, potentially reshaping how organizations approach vulnerability management and patching
Background
OpenAI has transitioned from a research lab to a commercial products company, offering enterprise services. Its Daybreak platform is the latest in a series of product launches aimed at the security market. The company's CEO and other executives have previously warned about catastrophic risks from advanced AI, including the potential for AI systems to enable large-scale cyberattacks. The lawsuit draws on those warnings as evidence that the companies themselves acknowledge the danger of continued development.
Key Perspectives
- OpenAI (Greg Brockman): Companies will be secure only to the extent of the frontier model they use to defend themselves and how well they utilize it; organizations need to reorient their security practice around what AI models can do.
- Nancy Wang, head of cyber-applied AI at OpenAI: Described Daybreak Red Team as delivering superhuman penetration testing at the scale of automated continuous scanning.
- Coalition of nonprofits (Irreplaceable Education Fund, Public Citizen, Heat Initiative, Alliance for Secure AI): Allege that AI companies have told the public their technology poses catastrophic risks, including enabling AI-driven cyberattacks and bioweapons, and that the companies are not wrong about those risks; seek a court order to stop development of AI more capable than current systems.
What to Watch
- How the San Francisco Superior Court rules on the nonprofits' request for an injunction
- Whether enterprise customers adopt OpenAI's security products or favor alternatives from other vendors
- Further regulatory or legal challenges to frontier AI development as companies continue to release more capable models
Checked
10 claims in this story checked against 1 source: 9 supported.
Show each claim
- SupportedOpenAI launched two Daybreak security products.www.theregister.com - Articles: “the company rolled out Daybreak Red Team and Daybreak Defense Factory, which are both part of OpenAI’s Daybreak security platform.”
- Partly supportedBrockman warned that companies’ security depends on using frontier AI models.www.theregister.com - Articles: “companies will be secure only to the extent of the frontier model that they use to defend themselves and how well they utilize that”
- SupportedA coalition of nonprofits sued in San Francisco Superior Court to stop OpenAI, Anthropic, Google, Meta and SpaceX’s AI businesses from developing or deploying systems more capable than their most advanced existing models.www.theregister.com - Articles: “a coalition of nonprofits filed a lawsuit in San Francisco Superior Court asking a judge to stop OpenAI, Anthropic, Google, Meta, and SpaceX's AI businesses from developing or deploying AI systems more capable than their most advanced existing models.”
- SupportedDaybreak Red Team uses OpenAI models to scan web applications for vulnerabilities.www.theregister.com - Articles: “Red Team, which is available in early access, uses OpenAI models to test companies’ web applications for vulnerabilities.”
- SupportedDaybreak Defense Factory finds vulnerabilities in codebases and develops patches for engineers to review and approve before production.www.theregister.com - Articles: “The agents scan organizations’ codebases, hunting for vulnerabilities and developing patches. Engineers review and approve the proposed fixes before they go into production.”
- SupportedBrockman said security fundamentals remain important but organizations need to reorient their security practice around what AI models can do.www.theregister.com - Articles: “security “fundamentals are important, and not to be discounted.” However, “you need to really reorient your security practice around what these models can do,” Brockman said”
- SupportedNancy Wang, who leads cyber-applied AI, described Red Team as superhuman penetration testing at the scale of automated continuous scanning, using agents with source code and application context to investigate attack paths.www.theregister.com - Articles: “Nancy Wang, who leads cyber-applied AI, described it as “superhuman pen-test at the scale of automated continuous scanning,” with a fleet of agents using source code and app context to investigate and test attack paths.”
- SupportedDaybreak Defense Factory uses agents built on OpenAI’s latest cyber models, currently GPT-6 Cyber.www.theregister.com - Articles: “This one also uses agents built on OpenAI’s latest cyber models - currently, that’s GPT-6 Cyber.”
- SupportedThe lawsuit cites incidents involving OpenAI agents accessing Hugging Face and unauthorized access to an Australian government portal containing Medicare statistics.www.theregister.com - Articles: “The lawsuit ... cites incidents involving OpenAI agents accessing Hugging Face, gaining unauthorized access to an Australian government portal containing Medicare statistics, and other AI-related cybersecurity incidents.”
- SupportedThe court documents allege that the companies say their technology poses catastrophic risks, including AI-driven cyberattacks and bioweapons, and that the companies are not wrong about those risks.www.theregister.com - Articles: ““Defendants tell us that the AI technology they are developing poses catastrophic risks to the public, including by enabling AI-driven cyberattacks and bioweapons,” the court documents allege, adding that the AI companies “are not wrong.””
