Remote job
AI Safety & Responsibility Policy Manager
Job details
About this role
Role overview Own and improve content policies that guide moderation across consumer products, enterprise services, and external model integrations. The role balances user safety, legitimate creative use, and operational practicality, translating policy decisions into effective machine-learning moderation and working with product and engineering partners to improve results.
Responsibilities - Maintain and evolve content policies as products, use cases, cultural norms, and market expectations change. - Translate policy requirements into prompts for large language model moderation and refine them to improve accuracy. - Audit moderation performance, identify recurring issues, and help resolve policy or system gaps. - Lead the policy contribution to red teaming, including maintaining automated test banks. - Build and update reference datasets used to measure moderation accuracy as policies and products evolve. - Turn audit and red-team findings into prioritized recommendations for policy, product, and engineering teams.
Requirements - Experience in content policy, trust and safety policy, or a closely related area. - Strong understanding of content moderation systems and hands-on experience using machine learning for moderation, policy enforcement, or risk assessment. - Excellent writing skills, including the ability to turn nuanced policy positions into clear, usable documentation. - Strong attention to detail and comfort reviewing large volumes of content, including potentially sensitive or disturbing material. - Ability to use data and statistics to inform policy decisions. - Initiative to identify and improve processes in a changing environment.
Nice to have - Familiarity with generative AI products, including video, image, or avatar applications. - Experience using large language models for policy drafting or enforcement, or building trust and safety infrastructure in a high-growth startup.
Benefits and work setup The role is open to remote candidates in the United States and Europe; offices are also available in New York City, San Francisco, Seattle, and London.