Uber’s Project Sandbox Dismisses AI Training Contractors

Uber has recently informed some of its artificial intelligence (AI) training contractors involved in Project Sandbox that their services are no longer required, terminating their contracts two months earlier than planned. This development was confirmed by various sources, including affected gig workers.

Overview of Project Sandbox

Project Sandbox is an initiative by Uber to facilitate AI training for other companies, specifically through collaborations with Google. This project is part of Uber’s broader ambition within its AI Solutions division. Reports indicate that approximately a dozen contractors were engaged in this project, although the exact number of terminations remains unclear.

Notification to Contractors

Uber communicated the decision via email, stating that a shift in client priorities impacted the ongoing work. The company reassured contractors that they would remain in the network for future opportunities that suit their profiles.

Background and Worker Qualifications

According to sources, many workers in Project Sandbox possessed advanced academic qualifications, such as PhDs, or had relevant industry experience. They were recruited earlier this fall and were initially promised at least three months of work.

AI Training Tasks

Contractors engaged in various tasks, including data annotation for photos and videos and evaluating AI-generated content for accuracy. Pay rates for these roles varied, with some contractors earning an hourly wage between $55 and $110, potentially leading to monthly earnings of around $19,000 at full capacity.

Recent Developments

• Uber’s stock has increased by approximately 39% so far in 2025.
• In the third quarter, Uber’s ride-hailing and delivery services reported significant growth in gross bookings.
• The company recently announced a pilot for its Digital Tasks program, which expands opportunities for gig workers to engage further in AI training.

Future Prospects for Affected Workers

Despite the abrupt termination, affected contractors are waiting for further instructions regarding the return of company hardware. Additionally, they have noted delays in receiving their first paychecks, which could take up to seven weeks from their start date.

Conclusion

The landscape for labeling and training AI continues to grow globally, with many participating as side hustlers or as primary income earners. The future of Uber’s AI training endeavors remains uncertain as the company navigates client priorities and broader market trends.