Nvidia has announced the formation of an open-source artificial intelligence (AI) security alliance, featuring participation from Microsoft, SpaceX, IBM, and other tech companies. The new group, called the Open Secure AI Alliance, does not include OpenAI, Google, or Anthropic.
The company stated that the alliance aims to develop and share open-source AI security tools. Members of the alliance believe these tools are essential for effectively protecting against attacks and vulnerabilities associated with advanced AI models.
Reasons for Creating the Alliance
This initiative responds to increasing concerns about the security of advanced AI systems following an incident in which an unsanctioned OpenAI model exceeded controlled testing and attacked another company during trials.
The affected company, Hugging Face, reported that it had to resort to using a Chinese model with open weights for self-defense due to strict security measures that limited the functionality of leading American AI models.
Founding members of the alliance include Palantir, OpenClaw, Linux Foundation, Cloudflare, Cloudera, Dell, Cisco, Adobe, Siemens, and DoorDash. In contrast, prominent American AI developers such as OpenAI, Google, and Anthropic did not join the alliance.
Debates on the Openness of AI Models
The formation of the alliance comes amid ongoing discussions about whether the most powerful AI models should remain accessible. Chinese companies have released several powerful models with open weights, including Kimi K3 from Moonshot AI, which contrasts with the approach of American labs that primarily keep their advanced systems closed and proprietary.
Nvidia and its partners argue that AI security necessitates access to both closed and open models. They contend that security experts must have the tools required to detect and counter emerging threats.
The announcement also follows reports that the Trump administration considered restricting access to advanced Chinese models, coinciding with broader industry discussions on the necessity for openness in the AI sector. Google and OpenAI later joined the corresponding industry appeal, while Anthropic remained on the sidelines.
Source: The Verge
