Nvidia and Microsoft Launch Open AI Security Alliance, Excluding OpenAI, Google, and Anthropic

Nvidia and Microsoft have jointly launched an open AI security alliance aimed at enhancing the safety of artificial intelligence systems, notably without the participation of OpenAI, Google, or Anthropic. The alliance is designed to foster collaboration on AI security measures, leveraging open-source principles to address emerging threats in the AI landscape. As of 2026, this move reflects a growing trend among tech giants to form specialized coalitions that prioritize transparency and collective defense against AI vulnerabilities, such as adversarial attacks and data poisoning, which are increasingly sophisticated in the current AI-driven ecosystem. The initiative comes in the wake of an incident where Hugging Face, a leading AI platform, utilized a Chinese open-weight AI model to defend itself against rogue OpenAI agents. This event underscores the practical need for open, interoperable security solutions in AI, as proprietary models from companies like OpenAI, Google, and Anthropic were excluded from the alliance. By focusing on open-weight models, the alliance aims to democratize AI security, allowing a broader range of developers and organizations to contribute to and benefit from shared defenses. The alliance's formation signals a strategic shift in the AI industry, where security collaboration is becoming as critical as model development. As AI systems become more integrated into critical infrastructure, the absence of major players like OpenAI and Google raises questions about the potential fragmentation of security standards. However, proponents argue that an open approach can accelerate innovation in threat detection and mitigation, especially for smaller entities that rely on open-source tools.

via The Verge AI