NVIDIA and Google have started working on AI security.

New Collaboration in AI Security

NVIDIA, Google, OpenAI, and AWS gathered in October 2025 with more than 40 representatives from government, industry, and academia to create the Center for Frontier AI Security (CFAS). This community aims to establish a new national center for collaboration at the intersection of artificial intelligence and national security.

CFAS has been established as an independent, non-profit organization dedicated to operationalizing artificial intelligence in the national security domain. Its goal is to bring together government, industry, academia, venture capital, and think tanks to raise awareness among leaders regarding the security of advanced AI models and to develop standards, tools, and mechanisms for national security applications.

Dr. Marina Theodotou, Founder and Executive Director of CFAS, stated, "CFAS aims to build a reliable and secure AI technology ecosystem in the U.S. by transforming existing policy research into actionable frameworks."

At the opening event, participants identified key priorities for ensuring the integration and interoperability of artificial intelligence in national security. These priorities included computing power, model assurance, verification testing, data management, and supply chain resilience. Leaders from various sectors emphasized the urgency of establishing shared standard frameworks and methodologies. It was highlighted that effective AI security requires both technical innovation and sustained public-private collaboration.

CFAS will serve as a think tank and implementation center that brings together expert working groups from government, industry, and academia to accelerate policy implementation, organize technical workshops, and conduct applied research collaborations.

CFAS aims to promote trust and collaboration between government agencies, researchers, investors, and technology providers, enhancing security, stability, and accountability while ensuring the integration and interoperability of AI. Additionally, it seeks to ensure that pioneering AI systems employed for national security purposes are secure, effective, and aligned with democratic values.