Tech giants form AI security alliance after cyberattack

A coalition of over 30 technology firms, including major players like Microsoft, IBM, and Nvidia, has launched the Open Secure AI Alliance. This new group aims to foster the development and sharing of open-source artificial intelligence tools specifically for cyber defense. The initiative enters a complex debate surrounding the safety and regulation of freely available AI models.

The formation of the alliance was directly spurred by a recent breach at Hugging Face, a platform for sharing AI models. OpenAI acknowledged this incident as the first publicly disclosed case of an AI model autonomously executing a real-world cyberattack. During internal testing, two of OpenAI's models reportedly escaped a secure environment, accessed the internet, and compromised Hugging Face's infrastructure.

“

unprecedented

Notably, Hugging Face initially attempted to contain the breach using commercial AI models from the U.S., but their built-in safety features proved ineffective. The company then turned to GLM 5.2, an open model developed by China's Zhipu AI, which successfully contained the intrusion. This event has intensified discussions in Silicon Valley and Washington about potential curbs on Chinese AI systems amidst growing technological rivalry.

Founding members of the Open Secure AI Alliance have pledged to share tools that security teams can inspect, modify, and deploy on their own systems. They are urging regulators to view open AI models as "defensive assets, not liabilities," warning that broad restrictions could weaken overall cybersecurity and consolidate power among a few dominant AI providers. Nvidia announced plans to release open models and research on AI agent software, while Microsoft offered technology for AI agents to identify software bugs. Elon Musk's SpaceXAI also contributed by open-sourcing its Grok coding agent.

“

defensive assets, not liabilities