The launch of the Coalition for Secure AI (CoSAI) on July 18, 2024, at the Aspen Security Forum marks a significant step towards establishing standardized practices for AI security. This initiative, hosted by the OASIS global standards body, brings together major tech industry players, academics, and experts to address the fragmented landscape of AI security.

CoSAI’s founding Premier Sponsors include Google, IBM, Intel, Microsoft, NVIDIA, and PayPal, with additional founding Sponsors such as Amazon, Anthropic, Cisco, Chainguard, Cohere, GenLab, OpenAI, and Wiz. This broad coalition of industry leaders demonstrates the widespread recognition of the need for secure AI development and deployment.

The initiative aims to provide practitioners and developers with guidance and tools to create Secure-by-Design AI systems. CoSAI’s scope encompasses the entire lifecycle of AI systems, including secure building, integrating, deploying, and operating. It focuses on mitigating various risks associated with AI, such as model theft, data poisoning, prompt injection, scaled abuse, and inference attacks.To achieve its goals, CoSAI will initially form three workstreams:

1. Software supply chain security for AI systems: Enhancing composition and provenance tracking to secure AI applications.
2. Preparing defenders for a changing cybersecurity landscape: Addressing investments and integration challenges in AI and classical systems.
3. AI security governance: Developing best practices and risk assessment frameworks for AI security.

The coalition is structured with a Project Governing Board that manages the overall technical agenda and a Technical Steering Committee composed of AI experts from academia and industry to oversee its workstreams.

CoSAI’s establishment addresses the current fragmented state of AI security, where developers struggle with inconsistent and siloed guidelines. By fostering a collaborative ecosystem to share open-source methodologies, standardized frameworks, and tools, CoSAI aims to democratize knowledge and advancements essential for secure AI integration and deployment.

The initiative welcomes contributions from all practitioners and developers as part of its open-source community, emphasizing inclusivity and collaboration. This approach is designed to accelerate the development of robust AI security standards and practices that will benefit the entire industry.As AI continues to transform various sectors, CoSAI’s efforts to establish standardized security practices and build trust among stakeholders globally are crucial. The coalition’s work is expected to have a significant impact on shaping a more secure and responsible AI future, addressing both classical and unique risks associated with AI systems. To read more, click the below icon.