Mercor, a $10 billion startup founded by three college dropouts in their early 20s, recruits highly skilled professionals to train AI systems by performing segmented tasks that teach models to replicate professional work across medicine, law, social work, and other fields.
The company has grown rapidly by exploiting economic anxiety—job openings in professional services have fallen over a million since 2022, and 42% of recent college graduates are underemployed, making Mercor's $90 average hourly pay attractive despite the risk of automating away their own careers.
Workers and contractors report significant problems including chaotic project management, surveillance through productivity-monitoring software, wage cuts without notice, strict NDAs preventing public complaint, and misclassification as independent contractors rather than employees—issues now being litigated in multiple class-action suits.
A March 2026 data breach exposed up to 4 terabytes of personal and training data through a compromised open-source tool, causing Meta to pause work with Mercor and prompting additional lawsuits over data protection failures.
Founders justify the AI replacement of white-collar work through techno-optimism—claiming productivity gains will create new jobs elsewhere—while evidence remains unclear whether AI will actually perform these professional tasks reliably or merely displace workers without creating equivalent opportunities.