Scale AI, Surge AI, and the billion-dollar gig-work industry shaping everything from chatbots to self-driving cars.

BY INC. VIDEO

Behind every major AI tool, including ChatGPT, is an enormous amount of human labor. Large language models don’t inherently know things. They must be trained, corrected, and fine-tuned by people. That work happens through a process called reinforcement learning through human feedback, or RLHF, which is essentially data labeling. Workers teach chatbots facts, correct their reasoning, flag mistakes, and help models learn how to respond appropriately. The same type of work guides autonomous vehicles, medical AI systems, search engines, and more. 

Like Uber drivers and DoorDashers, data labelers are gig workers. They’re paid hourly and receive no benefits. 

© Inc.com