AI Safety Data Reviewer (Japanese/English)

This remote, hourly-paid contract role involves reviewing AI-generated content and safety decisions, evaluating reasoning quality and step-by-step problem-solving, and providing expert feedback to ensure outputs are accurate, logical, and clearly explained. Responsibilities include assessing solutions for correctness and clarity, spotting methodological or conceptual errors, fact-checking as needed, and rating or comparing multiple responses based on safety and policy alignment. The work requires fluency in Japanese and the ability to make nuanced judgments across Japanese and English content. Your contributions help prevent the model from generating unintentional, adversarial, toxic, or unsafe outputs. Note: You may be exposed to potentially disturbing content, including sexual or violent topics, while supporting AI models for safe real-world deployment.