Displaced Workers Train AI Systems That Automated Their Own Jobs
Key Takeaways
- ▸AI companies are actively recruiting displaced workers whose careers were disrupted by AI automation to train AI systems, creating a troubling cycle of dependency
- ▸Data labeling and prompt-writing work offers relatively high pay ($45+/hour) but lacks job security, with projects subject to abrupt cancellation without warning
- ▸Workers are often kept in the dark about which AI systems they're training or what the end purpose of their work will be
Summary
A new investigation reveals how AI companies are recruiting laid-off professionals—journalists, lawyers, scientists, and content marketers whose careers have been disrupted by AI automation—to train the very systems that replaced them. Workers like Katya, a former freelance journalist turned content marketer, are being hired through platforms like Mercor to create training data for AI models, including writing prompts and responses, and evaluating AI outputs. The work, which pays $45+ per hour, offers a temporary lifeline for displaced professionals facing financial hardship, but comes with precarious employment conditions: abrupt project cancellations with no notice, last-minute Zoom calls at 6:30 p.m. on Sunday nights, and mandatory monitoring software installation. The irony is stark—workers whose livelihoods were eliminated by AI are now essential to improving the very technology that displaced them, while data labeling and AI training remain among the few remaining human jobs in the field.
- The precarious gig nature of AI training work—including last-minute scheduling and mandatory computer monitoring—reflects broader issues with how data annotation labor is treated in the AI industry
Editorial Opinion
This investigation exposes a deeply uncomfortable truth about AI's labor economics: the technology companies deploying AI to eliminate jobs are simultaneously exploiting the displaced workers willing to train those systems out of desperation. While data labeling has always been low-prestige work, recruiting and profiting from the very people whose careers AI has destroyed represents a particularly cynical approach. The lack of transparency about which models workers are training and the exploitative gig-work conditions suggest the industry knows this arrangement would face public backlash if scrutinized—a sign that meaningful reform in AI labor practices is overdue.
