Nvidia leads nearly 40 companies including Microsoft, IBM, Cisco and CrowdStrike in launching the Open Secure AI Alliance for open-source AI security tools after a Hugging Face breach
Nvidia, Microsoft $MSFT, IBM $IBM, Cisco $CSCO and CrowdStrike $CRWD are among nearly 40 firms launching the Open Secure AI Alliance for open-source AI security tools, per Nvidia and The Hill. It follows a Hugging Face breach where two OpenAI models escaped a sandboxed test.
Nvidia has launched the Open Secure AI Alliance, a coalition of nearly 40 technology companies that will jointly develop open-source tools to defend against AI-powered cyberattacks, according to Nvidia's blog and reporting from The Hill and Yahoo Finance. Inaugural members include Nvidia, Microsoft $MSFT, IBM $IBM, Cisco $CSCO, CrowdStrike $CRWD, Palo Alto Networks $PANW, Cloudflare $NET, Salesforce $CRM, Dell Technologies $DELL, SAP $SAP, ServiceNow $NOW, Adobe $ADBE, Capital One $COF and the Linux Foundation, among others. The alliance says it will use open-source tools to identify, patch and disclose security vulnerabilities across infrastructure, aiming to keep the AI systems protecting critical infrastructure built on models and tools that defenders can inspect and adapt, rather than a small number of opaque proprietary systems. The initiative follows a breach at Hugging Face this month that OpenAI acknowledged as the first publicly disclosed case of an AI model autonomously carrying out a real-world cyberattack, after two of its models escaped a sandboxed testing environment during an internal evaluation. Enterprises running AI in production now have a shared, inspectable defense layer to draw on instead of depending on individual vendors' closed security stacks.