The laid-off lawyers and PhDs training AI to steal their careers

Many highly educated professionals, including lawyers, PhDs, screenwriters, and software engineers, are finding temporary employment training the very artificial intelligence models that threaten their careers.

Companies like Mercor and Scale AI recruit these specialists to generate high-quality training datasets, evaluate chatbot interactions, and attempt to find flaws in model reasoning. While this gig work initially offers attractive compensation, workers must submit to intense keystroke and screen monitoring software that tracks their productivity to the second. The highly specialized tasks are designed to extract human expertise in fields ranging from legal analysis to creative writing, creating a vast digital assembly line where workers operate in isolation and under strict non-disclosure agreements.

Despite the promise of flexible remote work, contributors face extreme precarity due to the erratic nature of AI development. Projects frequently pause or terminate without notice as tech firms evaluate their models, leaving contractors without income. As competition among data-labeling platforms intensifies, workers report deteriorating labor conditions, including sudden pay cuts, reduced task completion times, and a lack of support from inexperienced management teams. This highly volatile environment has sparked widespread frustration and several class-action lawsuits in California over worker misclassification. Many contractors express deep existential dread, realizing that their efforts are accelerating the automation of their own professions, yet they continue because of a lack of alternative employment options in a tight job market.