It Turns Out That Google's AI Is Being Trained by an Army of Poorly Treated Human Grunts

Underpaid contractors train Google's Gemini AI chatbot under poor conditions, verifying accuracy and moderating content without proper warning. Despite portraying AI as autonomous, these systems rely on hidden human labor. Workers face tight deadlines, changing guidelines, and emotional distress while earning low wages. Google claims raters provide feedback but don't directly impact models, revealing the contradiction between AI marketed as labor-replacing and its dependence on exploited workers.