HackerOne AI Red Teaming is a service that provides adversarial testing for AI systems through human security researchers. The service tests AI models, APIs, and integrations to identify vulnerabilities including jailbreaks, misalignment, policy violations, and unsafe behavior. The service operates through time-boxed engagements tailored to specific threat models and risk priorities. Testing is conducted by AI security researchers from HackerOne's community who perform scenario-driven assessments aligned with frameworks including OWASP LLM Top 10, Gartner TRiSM, and NIST AI RMF. Engagements begin with scope definition and risk priority identification, followed by threat modeling and test plan development. Researchers conduct adversarial testing to uncover vulnerabilities that automated tools may miss. Results are delivered through the HackerOne Platform with detailed findings, prioritized recommendations, and remediation guidance. The service includes collaboration with solution architects who assist with attack scenario design, fix validation, and mitigation acceleration. Testing covers full-stack AI deployments including models, applications, and system integrations. Organizations receive centralized reporting for tracking and managing remediation efforts.