Workers are paid to train artificial intelligence systems (1) to think more like humans, and in some cases, they are teaching machines how to do the very jobs they once feared AI would replace.
That’s what happened to Hollywood writer and showrunner Ruth Fowler. In 2023, entertainment workers (2) went on strike in part over fears that studios could use AI to replace writers and actors. But after the strike ended, work did not fully resume. When another producer defaulted on a six-figure payment he was owed, Fowler found himself looking for a way to stay afloat.
Must read
“I was looking for easy money. I also needed cash to pay rent and buy food,” Fowler wrote in an essay for Wired (3). “How hard could it be to teach a machine to take my job? I was naïve enough to believe that this industry wanted what we had to offer: not just our skills, but us.”
But it wasn’t just the writers. Companies are hiring lawyers, doctors, venture capitalists, coders and foreign language speakers to help train artificial intelligence systems.
A new type of secondary activity
A company that leans towards this trend is Mercor (4), whose pitch to workers is simple: “get paid to work on AI projects.” A current listing on its Physician Talent Network (5) advertises paying up to $250 an hour to doctors who help train AI systems through medical scenarios, answer reviews, and expert feedback.
And experts say demand for these roles is expected to grow as artificial intelligence systems evolve. As many large language models have already been trained with large amounts of existing online data, the next phase of development increasingly relies on human input to fine-tune responses, improve accuracy, and help systems perform better in specialized areas.
Mercor CEO Brendan Foody told CBS News (6) that the company wants expertise in almost every field.
“We hire everyone from chess champions to wine aficionados to help train (AI) agents to be better, because ultimately we want them to know how to give better advice in a chess game or recommend what wine to have at dinner,” he said.
Hollywood writer Robin Palmer said she now spends about 30 hours a week helping train AI through projects with Mercor, evaluating whether the technology can produce stronger, more compelling creative writing.
“They’re turning in work and you’re wondering, ‘Does it work structurally? What’s the characterization like? Are there clunky transitions?'” he told CBS News (7). “I really like seeing how AI is improving. It’s almost like working with a student and saying, ‘Yes, you’re improving.'”
Read more: Non-millionaires can now accumulate properties like those of the 1%: how to start with just $100
The fine print of AI work
For Fowler, the daily reality of work was very different. One of their first tasks involved reviewing conversations between users and AI chatbots, rating how the systems responded to deeply personal questions, and rating the answers on a scale of one to five.
But the flexibility and promise of easy money came with a reality check. Fowler recalled receiving a late-night Slack message from a team leader warning him not to trust the work.
“These aren’t jobs,” he recalled Fowler being told. “They are “tasks” and we are “tasks.”
That uncertainty may be one reason why many workers remain uneasy about the growing role of AI in the workplace. While these projects are creating new ways for some people to make money, a recent Pew Research Center survey (8) found that more than half of employees are concerned about the long-term impact of AI on work, while nearly a third believe the technology could eventually reduce job opportunities in the coming years.
Opportunity or warning sign
Palmer acknowledged that some in Hollywood may consider working with AI controversial, but he said he believes experienced professionals can help shape the technology responsibly, while acknowledging that the growing presence of AI in the workplace may be difficult to avoid.
“The train has left the station,” he said. “So, do you want AI to be good because it’s being trained by good people, or not?”
AI training has become an unexpected source of income for some workers and a way to stay relevant as industries change rapidly. Others see this as raising uncomfortable questions about whether they are helping to create tools that could eventually reduce demand for their own skills.
Fowler landed firmly in the second camp. After trying to make a living in the emerging AI economy, he wrote that the experience was “crueler than I ever imagined.”
“They will be tasked with making us work faster and longer, with more precision, more control, fewer errors, less overhead, less cost. To make the machine more human, they will make us more like the machine,” he wrote.
That tension may ultimately define the next phase of AI in the workplace: Some see an opportunity to adapt and profit from a rapidly growing industry, while others feel like they are training a replacement before fully understanding what comes next.
you may also like
Join over 250,000 readers and get the best Moneywise exclusive stories and interviews first – clear insights curated and delivered weekly. Subscribe now.
Article sources
We rely only on verified sources and credible third-party reports. For more details, see our ethics and guidelines.
CBS News (1), (6), (7); Los Angeles Times (2); Wiring (3); Mercor (4), (5); Pew Research Center (8)
This article originally appeared on Moneywise.com with the title: ‘The train has left the station’: Workers are cashing in by teaching AI to do its job; some earn up to $350 an hour
This article provides information only and should not be construed as advice. It is provided without warranty of any kind.
