Gray Swan AI is an AI security company that tests models and AI agents for vulnerabilities and misuse. Its products are Shade, an automated adversarial testing tool, and Cygnal, a runtime solution that classifies and blocks adversarial inputs and unsafe agent outputs. The company also runs Arena, a red-teaming network, and provides adversarial evaluations and red-teaming engagements for model builders and for organisations deploying AI. The services are delivered as SaaS and as project-based work.
The company is Gray Swan Security, Inc., based in Pittsburgh, United States, and a spin-out of Carnegie Mellon University. It is therefore subject to US law, including the CLOUD Act. Its privacy policy makes no GDPR commitment and names neither a hosting region nor a data processing agreement; its Trust Center lists SOC 2 Type 2 and Cyber Essentials. Where customer data is hosted is not publicly stated. European organisations that share test data or model behaviour with the service should have this confirmed contractually beforehand.