Open Secure AI Alliance: NVIDIA Leads AI Security Initiative

A new industry-led open-source initiative, the Open Secure AI Alliance, is forming to secure AI models and digital infrastructure. NVIDIA announced this consortium, which includes 37 founding organizations, to make AI safer and more secure through open-source technologies. The alliance builds on existing initiatives and plans to develop tools for finding vulnerabilities and auditing AI systems. Founding members include major tech companies like Adobe, Microsoft, IBM, and SpaceXAI. They will contribute open models, datasets, and security tools. For example, NVIDIA is contributing NOOA, an open platform for testing AI agent behavior, and Microsoft is offering MDASH for multi-agent vulnerability discovery. SpaceXAI is adding its Grok Build coding agent. The creation of this alliance was partly spurred by a recent security incident that highlighted the limitations of closed AI systems in cyber investigations. The bottom line is, this collaboration aims to boost trust and transparency in AI by making security tools openly accessible and collectively developed.