Professionals earn up to $350 an hour training AI amid job security concerns

Professionals including writers, doctors, and lawyers are increasingly earning income by training artificial intelligence systems, with rates reaching up to $350 an hour. This trend is largely driven by companies like Mercor, which recruits experts to fine-tune AI responses and provide specialised feedback. For some, including Hollywood writer Ruth Fowler and Robin Palmer, this work emerged as a financial lifeline following industry instability and the 2023 Hollywood strike.

Mercor’s Physician Talent Network advertises pay of up to $250 an hour for doctors reviewing medical scenarios and providing expert feedback. The company’s CEO, Brendan Foody, stated that the firm hires experts from diverse fields, including chess champions and wine hobbyists, to help AI agents provide better advice in specific domains. As large language models have already been trained on vast amounts of existing online information, the current phase of AI development relies more heavily on human input for fine-tuning.

Ruth Fowler, a Hollywood writer and showrunner, participated in AI training tasks after facing financial difficulties, including a defaulted six-figure payment from a producer. She described her experience as "more cruel than I could have ever imagined," noting that workers are being tasked with making machines more human while becoming more like machines themselves. Fowler recalled being told by a team leader that these were not jobs but "tasks," and the workers were "taskers," highlighting the precarious nature of the work.

Robin Palmer spends approximately 30 hours a week evaluating AI-generated creative writing for structural integrity and characterisation. While she views the work as a way to shape AI responsibly, she acknowledges the tension inherent in the role. "The train has left the station," Palmer said, questioning whether AI should be trained by good people to ensure it performs well.

A recent survey by the Pew Research Center found that more than half of employees are concerned about AI’s long-term impact on work, and nearly one-third believe it could reduce job opportunities in the coming years. This anxiety is compounded by the 2023 Hollywood strike, where entertainment workers went on strike partly due to fears that studios would use AI to replace writers and actors.

The rise of AI training roles presents a paradox for the labour market. While it offers immediate financial relief for professionals facing industry shifts, it also raises uncomfortable questions about whether they are helping to develop technology that may eventually displace their own roles. The trend underscores a broader shift in how expertise is valued and compensated in an era of rapid technological change.

As the demand for human input to refine AI systems grows, the line between worker and trainer continues to blur. Professionals are adapting to a new reality where their skills are essential for current development but may be rendered obsolete by the very systems they help create. This dynamic is likely to define the next phase of the labour market, where job security and technological adaptation remain closely linked.