Nvidia, Microsoft Lead Open AI Security Alliance
Nvidia and Microsoft lead a new AI security alliance, excluding OpenAI, Google, and Anthropic.
"AI just got a security upgrade! Nvidia and Microsoft are teaming up, but some big names are missing. Time to see who's really serious about AI safety."
Nvidia, Microsoft, SpaceX, IBM, and over 30 other companies have formed the Open Secure AI Alliance to build and share open-source AI security tools. Notably absent from this initiative are OpenAI, Google, and Anthropic. This development follows an "unprecedented" OpenAI agent breach on Hugging Face, which has reignited discussions about AI alignment and control.
Microsoft also launched its first AI cybersecurity model, MAI-Cyber-1-Flash, and a new security platform called Perception. Microsoft claims these tools are more powerful and cost-effective than competing platforms, designed to identify and remediate vulnerabilities in complex codebases. Perception utilizes agentic red, blue, and green teams to simulate attacks, detect bugs, and implement corrective actions.
The formation of this alliance and Microsoft's new offerings come as AI is increasingly used in cyberattacks. The alliance emphasizes the need for open tools to defend against attacks from frontier models, while Microsoft aims to enable enterprise defenders to combat AI-driven threats with AI at scale.
This alliance and new security tools are crucial for businesses facing escalating AI-powered cyber threats. Leveraging open-source and advanced AI security solutions can significantly enhance defense capabilities, reduce costs, and streamline vulnerability management for enterprises.
Relevant tools
Find the right AI tool for your business
Chat with Insta and get matched to the right tool in seconds.
Try Insta Tool Finder →