The Open Secure AI Alliance builds and shares an open defensive stack of AI models, tools, and techniques to strengthen cyber defense and secure software and AI agents—giving defenders everywhere capabilities they can inspect, adapt, trust, and control.
Open source earned its role in the modern economy not by being free, but by being observable. The same principle that secured the internet now needs to secure AI. Open systems provide the transparency, adaptability, local control, and flexibility that security teams need to strengthen their environments.
Organizations need to understand and govern the full agent stack. Open systems give defenders the ability to inspect, adapt, and run advanced capabilities on infrastructure they control.
Organizations can address AI-era threats better together. The Alliance brings defenders together to test systems, share evidence, and turn discoveries into protections the entire ecosystem can use. When many organizations can build on and adapt capable models, rivalry extends beyond model developers to chips, applications, and services — keeping AI’s benefits broadly shared.
Portable defenses remain effective as models, vendors, and environments change. They provide flexibility and diverse solutions while expanding innovation across models, infrastructure, applications, and security services.
An AI agent is more than a language model. It’s a software system built from models, the harnesses that let it observe context and take action, and the guardrails that constrain what it can do.
Much of the AI safety conversation has focused on the model alone. But security depends on the full system: models, harnesses, alignment mechanisms, runtimes, identity, policy, enforcement points, observability, and recovery. These elements need to be just as open, testable, and auditable.
SAFE confidentially collects and analyzes AI incidents and near misses, promptly informs affected parties and turns recurring failures into shared, evidence-based controls that reduce systemic risk.
Leaders across cloud computing, cybersecurity, enterprise software, open source foundations, and AI research have joined as inaugural partners.
The Open Secure AI Alliance is open to industry, researchers, and government partners who want to build the defensive infrastructure the AI era requires.
✓ Contribute open tools, models, or research to the shared defense stack
✓ Collaborate with foundations like OpenSSF on shared standards
✓ Help shape guidance for policymakers and regulators
