After graduating from college in Nairobi, Joe found work as a data annotator, labeling images and data to be used for training artificial intelligence (AI). He later established an annotation boot camp for a company, Remotasks, producing hundreds of trained annotators, despite the difficult and monotonous nature of the job. The company is a subsidiary of Scale AI, a major data vendor with clients including OpenAI and the U.S. military, information unbeknownst to Joe and his students. These human annotators play a critical role in the AI supply chain, with their often unrecognized labor being foundational for AI’s capabilities and progress.
- Joe’s job involved the intensive labor of tagging and categorizing data, a foundational aspect of training AI systems. He later ran an annotation boot camp, producing many trained annotators.
- The work, although essential, was mundane and monotonous, causing many of Joe’s students to drop out before the boot camp was finished. The tasks they were working on were often small components of larger projects, making the overall purpose unclear to the workers.
- Remotasks, the company Joe worked for, is a subsidiary of Scale AI, a multi-billion dollar data vendor, serving prominent clients like OpenAI and the U.S. military. The connection between Remotasks and Scale AI was unknown to the workers.
- Despite AI’s growing sophistication, human labor remains essential in labeling and clarifying data, especially for ‘edge cases’ where AI might fail. However, this labor often remains hidden and undervalued.
- The AI annotation industry, despite its significance, is shrouded in secrecy and complexity, often unbeknownst to the very people working in it. As AI continues to evolve, the need for human annotators is likely to persist, requiring a shift in recognizing and valifying these roles.