NVIDIA gathers the industry chain to establish the "Open and Secure AI Alliance" and supports the open weight model
As AI agents develop rapidly and network security risks continue to rise, NVIDIA is trying to unite industry chain partners to build an open and secure AI alliance. On Monday local time, Nvidia announced the establishment of the OpenSecureAIAlliance (OSAA), which will unite Adobe, Dell Technologies, CrowdStrike, HuggingFace and other technology companies to jointly develop and share AI security and network security tools, hoping to build a more complete AI security ecosystem with the help of open models.
As AI agents develop rapidly and network security risks continue to rise, NVIDIA is trying to unite industry chain partners to build an open and secure AI alliance. On Monday local time, Nvidia announced the establishment of the Open Secure AI Alliance (OSAA), which will join forces with Adobe, Dell Technologies, CrowdStrike, Hugging Face and other technology companies to jointly develop and share AI security and network security tools, hoping to build a more complete AI security ecosystem with the help of open models. NVIDIA stated in a statement that as governments and regulatory agencies continue to improve the AI security supervision system, open models, open agent operating frameworks, and open security tools should be regarded as important defensive assets for AI security and network security, rather than potential risks. The background for the establishment of this alliance is an AI security incident recently disclosed by OpenAI. According to public information, an experimental AI agent once escaped from preset control and launched a network intrusion into the AI development platform Hugging Face. According to reports, the agent was discovered by OpenAI after the attack continued for a period of time. The incident was eventually brought under control and U.S. regulatory agencies were notified. Although the incident did not cause more serious consequences, the industry generally believes that this means that AI agents have stronger autonomous execution capabilities, and it also indicates that network offense and defense are entering a new stage of AI versus AI. Nvidia said in a statement that recent events have once again proven that network defense requires advanced AI agents that can be deployed independently and quickly adjusted. If security teams cannot run, inspect, and optimize advanced AI models on their own infrastructure, their ability to respond to a security incident will be significantly limited when a security incident actually occurs. To this end, OSAA will focus on cooperation in vulnerability discovery, vulnerability disclosure, security assessment, and AI agent governance in the future, and share security capabilities through open technology to promote the establishment of a unified AI security infrastructure. At the same time, Nvidia also announced that it will open models, model weights, data sets, and AI agent research results to the alliance, including the latest open source project Nvidia Labs Object-Oriented Agent, to accelerate the development of a new generation of network security tools and technologies. This framework is currently available on GitHub, which can help developers test, track, review and manage AI agent behavior more conveniently, and improve the controllability of complex AI systems. It is worth noting that the establishment of OSAA also continues the recent US technology industry’s support for open weight AI models. Just a few days ago, a number of technology companies, including NVIDIA, signed an open letter calling on policymakers to support the development of open-weighted AI models and avoid prematurely introducing restrictive measures to avoid weakening innovation capabilities and over-concentrating AI capabilities on a few closed platforms. Nvidia believes that implementing "one-size-fits-all" restrictions on open cutting-edge AI systems will not only weaken network defense capabilities, but may also lead to further concentration of AI capabilities, security capabilities, and infrastructure on a few closed-source providers, increasing the dependence and systemic risks of the entire industry. The company said that an open ecosystem means that more research institutions, security teams and enterprises can continuously review models, discover vulnerabilities and jointly improve security mechanisms, thereby improving the security level of the entire industry rather than weakening security. From the perspective of alliance members, OSAA covers many key links in the AI industry chain. In addition to the above-mentioned companies, it also includes dozens of technology companies such as Microsoft, SpaceX, and Palantir, covering many fields such as AI models, cloud computing, network security, software, and hardware. For Nvidia, this promotion of the establishment of the Open and Secure AI Alliance also means that it is further extending from an AI computing power provider to an AI software ecosystem builder. As more and more companies begin to deploy AI agents, how to establish a reliable, safe, and controllable governance system while improving model capabilities will become an important direction for the development of the AI industry in the next stage.