Uber's AI Training Gig: Short-Lived Opportunities

Uber has recently let go of a group of AI training contractors, ending their work two months earlier than initially promised. These workers were part of Project Sandbox, a program where Uber provides AI training services to other companies, specifically Google in this case. Around a dozen contractors were involved, though the exact number affected is unclear. The workers, who had advanced degrees or relevant experience, were told that their services were no longer needed due to a change in the client's priorities. Uber assured them that they would be considered for future opportunities. However, this sudden change left many workers in a tough spot, especially since they were still waiting for their first paycheck, which Uber stated could take up to seven weeks to process. The tasks these contractors performed included annotating photos and videos, as well as evaluating AI-generated responses. Pay rates varied, with some workers earning up to $110 an hour if they consistently worked 40 hours a week. However, they were sometimes asked to work fewer hours, making it difficult to reach the highest pay rate. This situation highlights the unstable nature of gig work, particularly in the AI training sector. While Uber has been successful in its ride-hailing and delivery services, its AI training arm is still in its early stages. The company recently launched a new program called Digital Tasks, which allows gig workers to train AI. Some of these tasks require advanced degrees, showing Uber's expansion beyond driving and delivery. The workers involved in Project Sandbox were recruited through a staffing agency and were promised at least three months of work. They were also provided with company laptops, which they now have to return. This experience has left some workers questioning the reliability of gig work, especially in the tech industry.