Aspen, CO – In a landmark initiative at the Aspen Security Forum, Google announced the launch of the Coalition for Secure AI (CoSAI). This industry forum, designed to enhance AI security, brings together key players in technology and academia to address the unique risks posed by AI development and deployment.
The founding members of CoSAI include Amazon, Anthropic, Chainguard, Cisco, Cohere, GenLab, IBM, Intel, Microsoft, NVIDIA, OpenAI, PayPal, and Wiz. This coalition will operate under the umbrella of OASIS Open, an international standards and open-source consortium.
CoSAI’s Initial Focus Areas
CoSAI has outlined three primary areas of focus for its inaugural efforts:
- Software Supply Chain Security for AI Systems: Leveraging Google’s SLSA Provenance framework, this initiative aims to enhance AI software security by tracking and verifying the entire lifecycle of AI models. This includes assessing third-party risks and expanding existing SSDF and SLSA principles to cover AI applications comprehensively.
- Preparing Defenders for a Changing Cybersecurity Landscape: This workstream will create a defender’s framework to aid security practitioners in navigating AI-related security challenges. It aims to scale mitigation strategies to keep pace with offensive cybersecurity advancements within AI models.
- AI Security Governance: Developing a comprehensive taxonomy of AI security risks and controls, this initiative will provide resources such as checklists and scorecards to help organizations assess, manage, and monitor the security of their AI products effectively.
Collaboration and Future Goals
CoSAI will collaborate with various organizations, including the Frontier Model Forum, Partnership on AI, Open Source Security Foundation, and ML Commons, to promote responsible AI practices.
Heather Adkins, VP of Security Engineering at Google, and Phil Venables, Vice President and Chief Information Security Officer (CISO) of Google Cloud, emphasized the importance of a collective investment in AI security. As AI technologies evolve, CoSAI aims to ensure that risk management strategies advance accordingly, providing a robust framework for secure AI implementation.
Learn More and Get Involved
For more information on CoSAI and how to support its initiatives, visit coalitionforsecureai.org. To learn more about Google’s Secure AI Framework, visit the Secure AI Framework page.
