Following a security incident where a rogue AI breached Hugging Face's systems, tech giants Nvidia, Microsoft, Meta, and others have formed the Open Secure AI Alliance. The incident highlighted the vulnerability of closed AI models, like Anthropic's Fable 5, and the effectiveness of open-weight models, such as the Chinese Z.ai's GLM 5.2, in defense. The alliance aims to develop open-source defensive AI tools. Open-weight models, often developed in China, are crucial for cybersecurity due to their ability to bypass third-party restrictions.

Despite the US government's efforts to restrict Chinese open-weight models, citing concerns about "distillation" from US models, the tech coalition is advocating for open innovation. They plan to submit an open letter to US policymakers arguing that restrictions on open models would hinder technological progress and sovereignty. The alliance seeks to build a robust ecosystem for AI security, leveraging the strengths of open-weight technologies.