5 people explain how they broke into AI training and how much they make in their side hustle (www.businessinsider.com)

Five contractors told Business Insider how they broke into AI training work — a booming, decentralized side hustle that helps large companies (Meta, Google, OpenAI) and startups improve chatbots and models. Platforms named include Prolific, Outlier/Remotasks (Scale AI), Mercor, Mechanical Turk and others. Tasks range from rating AI responses, fact-checking and multimedia prompt engineering to annotating videos and training voice/chat models. Contributors — from teachers and Ph.D. scientists to e‑commerce entrepreneurs — report flexible schedules and meaningful pay: examples include monthly take‑homes of $1,000–$1,200, $20–$100/day or roughly $20/hr typical, specialist projects or advertised rates up to $100/hr, and one contributor who earned $31,000 over 18 months. The story highlights why human-in-the-loop work is critical as public datasets age: companies are hiring thousands of labelers and domain experts to preserve model quality and native-language coverage. But the gig has downsides — long waitlists, churny short-term projects, task-time underestimation that reduces effective hourly rates, tedious workflows and uneven management — even as it builds valuable skills (QA, prompt engineering, domain expertise). For the AI/ML community this signals continued reliance on distributed, often under‑recognized human labor for model safety and performance, raising implications for workforce standards, pay transparency and annotation quality.