Mercor Hires Doctors and Lawyers to Train AI Systems to Automate Their Own Specialized Work

A new Silicon Valley startup is putting into practice what AI executives have long warned about in theory: systematically training artificial intelligence to replace skilled professionals across multiple industries. Mercor, a San Francisco-based company founded in 2023 and valued at around $10 billion, is hiring doctors, lawyers, bankers, and other highly trained professionals to teach AI systems how to perform their own specialized work—raising urgent questions about the gap between public rhetoric and private action in the AI industry.

The company's business model relies on a process called reinforcement learning with human feedback. Mercor employs domain experts to review responses generated by large language models like ChatGPT, Claude, and Gemini, then correct errors, rank outputs, and provide structured feedback. A doctor evaluates medical recommendations, a lawyer examines legal documents, and a banker assesses financial analysis. This human feedback is fed back into the AI system, allowing it to refine its reasoning and improve its accuracy in specialized fields. In essence, the startup is paying experts to document their expertise so machines can eventually replicate it without human involvement.

The timing is notable given that many AI company leaders have become increasingly vocal about job displacement concerns. OpenAI CEO Sam Altman and executives at other major AI labs have publicly acknowledged that widespread job loss from artificial intelligence is likely. However, according to observers, there appears to be a disconnect between these cautionary statements and the actual business strategies being pursued by companies in the AI space. While executives warn about potential displacement, startups like Mercor are actively working to accelerate it, building the infrastructure to automate high-skill professional work that traditionally requires years of education and experience.

The startup was founded by three college dropouts—Brendan Foody, Adarsh Hiremath, and Surya Midha—who notably had never held traditional professional jobs themselves. This background may be significant, as they appear unconcerned about the employment implications of their work. The company's rapid valuation reflects investor confidence in the market potential of AI automation, suggesting that despite public concerns about job loss, significant capital continues flowing toward companies explicitly designed to replace human workers in lucrative professional fields.

The implications extend beyond individual job losses. Industry observers have raised the possibility that AI could displace between one and two billion jobs globally over the next decade, which would represent an unprecedented economic disruption. The existing labor market infrastructure lacks capacity to retrain and reskill such a displaced workforce, creating what some describe as a fundamental mismatch between the pace of technological change and society's ability to adapt to it.

What remains unclear is whether the public acknowledgment of job loss risks by AI executives represents genuine concern or strategic positioning. By openly discussing displacement while their companies simultaneously build the systems to cause it, AI leaders may be attempting to shape policy conversations and manage public perception. For workers in professional fields, the message from Mercor's rise is unmistakable: the theoretical warnings about AI replacing skilled work are already becoming practical reality.