Gray Swan

Gray Swan provides safety and security solutions for AI deployments. The company focuses on addressing threats faced by AI systems in environments where well-resourced attackers, external tool use, retrieval, and misuse could compromise operations. Their offerings include tools that assess the risks of AI deployments and secure models designed to provide security and safety capabilities. The company serves enterprise organizations, frontier model developers, startups, and other entities that need to deploy AI systems with confidence. Gray Swan's approach addresses vulnerabilities in large language models and other AI systems that could be exploited through various attack vectors. Gray Swan was founded by experts in AI safety and security from Carnegie Mellon University, where the founders conducted research on identifying vulnerabilities in large language models. This research has contributed to understanding how to build safer AI systems and has been featured in major media outlets including the New York Times, The Washington Post, TIME, and CNN. The company's founders have published papers at machine learning and AI conferences, with several receiving paper awards. The company's focus areas include vulnerability assessment for AI deployments, securing AI models against attacks, and providing security solutions tailored to the unique challenges of AI systems in production environments.