Workers are getting paid to teach AI how to do their jobs

Leading AI companies are actively recruiting individuals with diverse expertise to train their generative AI models, creating a surge in specialized jobs. According to job postings, professionals such as Hollywood screenwriters, air traffic controllers, litigators, and even wine hobbyists are being hired to refine AI responses in their respective fields. Christine Cruzvergara, vice president of higher education and student success at Handshake, notes these roles are among the fastest-growing in the industry as large language models require more fine-tuning. Hollywood screenwriter Robin Palmer dedicates 30 hours weekly to teaching chatbots how to produce compelling creative writing, comparing the process to mentoring a fledgling writer. AI labs prioritize hiring experts—like chess champions or psychiatrists—to improve model accuracy in niche areas. For instance, psychiatrists earn up to $350 an hour to design clinical scenarios and evaluate AI outputs in mental health care. Mercor, a recruitment platform for AI companies, reports these training jobs pay an average of $105 per hour, with some roles requiring non-disclosure agreements. Generalists, who review and annotate AI search results, earn $50 hourly and need strong communication and reasoning skills. Brendan Foody, CEO of Mercor, predicts training AI agents will become the world’s largest job category. While some workers fear AI will eventually replace them, Palmer acknowledges the inevitability of AI integration, even if it risks alienating colleagues. The 2023 Hollywood strikes highlighted industry-wide concerns about AI displacing jobs, from actors to animators. Despite skepticism, experts like Foody emphasize the value of human guidance in shaping AI to serve society effectively. Job listings often conceal the hiring AI developers, but roles span creative writing, technical fields, and specialized domains. The demand reflects AI’s need for human expertise to refine outputs, ensuring accuracy and relevance across industries.