This Silicon Valley startup is hiring experts to train AI to replace doctors, lawyers, bankers and other jobs
A new Silicon Valley startup is hiring skilled professionals like doctors, lawyers, and bankers to train AI systems like ChatGPT and Cloud to handle high-skilled professional work. The startup aims to help larger language models improve in particular areas.
A new startup in Silicon Valley is paying humans to teach AI their jobs, so that one day machines can do the same jobs themselves. And no, this is not the plot of some science-fiction experiment, but the business model of a startup called Mercor, which is already valued at about $10 billion. The company is hiring highly skilled professionals to train artificial intelligence systems for high-value tasks that traditionally require years of education and experience.
❮❯
Mercor is a San Francisco-based startup founded in 2023 by three friends and college dropouts, Brendan Foody, Adarsh Hiremath, and Surya Midha. The company hires domain-specific professionals to train AI models. Right now, big language models like Gemini, Cloud, and ChatGPT are powerful, but they don’t automatically understand medical nuances, legal reasoning, or financial risk. To help improve these AI systems, this startup works with professionals to review AI-generated responses, fix errors, rank the outputs, and provide structured feedback.
This process is called reinforcement learning with human feedback. This process enables these machine learning systems to refine their logic step by step and become more reliable in particular areas.
Simply put, this startup hires real experts to test how AI works. A doctor reviews its medical recommendations, a lawyer examines its legal draft, and a banker evaluates its financial analysis. Their improvements and ratings are fed back into the system, helping the AI learn from its mistakes and get better over time.
Although the startup is clearly making money by helping improve large language models, there’s also an uneasy feeling to its rise. Many critics argue that in reality, the professionals are training their replacements themselves. If AI becomes capable of handling routine research, analysis and documentation in fields such as law, finance and medicine, entry level and support roles could be significantly reduced.
However, many professionals involved in the process see it differently. Doctors working with the startup argue that medication is about more than pattern recognition. This includes empathy, intuition and meaningful patient interactions, qualities that machines cannot easily replicate. From this perspective, AI can take over the administrative burden and information-heavy tasks, freeing up doctors to focus on the human side of care.
“I don’t want to look at it as AI taking over our jobs. I see it as AI taking over aspects of our jobs that prevent us from being good doctors, good therapists, and good listeners,” Dr. Alice Chiao, who previously taught emergency medicine at Stanford University’s medical school, told CNN.
Still, fear cannot be ignored. Many in the tech industry believe it is only a matter of years or even months before AI systems become significantly more capable than human labor. XAI CEO Elon Musk previously predicted that AI could be smarter than individual humans by 2026 and potentially overtake humanity collectively by the end of the decade. This change will likely cause widespread disruption, especially in entry-level roles.
In fact, early signs of this change are already visible. A recent analysis from Ireland’s Department of Finance warns that AI is already disrupting entry-level roles in the finance and technology sectors in the country, and it could soon become a global reality.
Even some pioneers of AI have acknowledged that AI will bring disruption in the future. Geoffrey Hinton, often called one of the “godfathers of AI”, has previously said that the technology will significantly reshape employment, especially in sectors that rely heavily on administrative and analytical tasks. Recently, Microsoft’s AI head Mustafa Suleiman also warned that AI could automate most of the white-collar tasks within the next 12 to 18 months.
