AI Won’t Just Take Jobs. It Will Change What a Job Means
Will artificial intelligence replace us—or simply change the way we work?
That question is no longer theoretical. It is becoming a workplace reality.
Music producer Robbie Hiser is one of more than 100,000 freelancers working with Mercor to train AI models on everything from finance to poetry. He teaches machines some of the skills he has spent his career developing.
Yet Hiser isn't convinced he's training himself out of a job.
He argues that great creative work still requires something AI doesn't possess: a life.
“AI doesn't have any life experience to pull from,” he says. “An AI never had a girlfriend break up with it.”
It's a funny line. But it points to a serious question: What happens when machines become very good at the work we've traditionally considered uniquely human?
The optimistic view: AI as a co-worker
Mercor CEO Brendan Foody sees AI as the next step in workplace evolution.
The argument is familiar: let machines handle the repetitive, tedious work, and humans can spend more time on strategy, creativity, judgment and relationships.
Some companies are already experimenting with that model.
A Columbus law firm, for example, worked with Stanford University's Liftlab to create AI personas based on the expertise and philosophy of senior lawyers. Those systems can help attorneys review documents and prepare cases.
But they aren't supposed to replace the lawyers.
As attorney Kim Herlihy puts it, AI may help prepare the courtroom argument—but it isn't going to stand before a jury and deliver it.
And that's the optimistic vision of the AI economy: machines do the work around the work, while humans remain responsible for the moments that matter.
The darker possibility: fewer people doing more
Economist Daron Acemoglu isn't nearly as optimistic.
He argues that the scale and speed of AI automation could be fundamentally different from previous technological revolutions.
The Industrial Revolution unfolded over decades. AI is advancing across multiple industries simultaneously—and in years rather than generations.
The result could be enormous productivity gains.
It could also mean enormous displacement.
Acemoglu warns that, under a worst-case scenario, unemployment could triple over the next decade.
And there's an uncomfortable mathematical problem with the idea that displaced workers can simply find jobs training AI.
AI training may be growing rapidly and can pay surprisingly well. But the number of people needed to train AI is tiny compared with the number of workers whose tasks AI could eventually automate.
You can't solve mass automation by hiring everyone to automate everyone else.
The workers feeling it first
The impact is already showing up among younger workers.
Research cited by CBS indicates that hiring has weakened among young people in some occupations most exposed to AI, including software development. Other research has found declines in hiring and wages among recent graduates in AI-exposed fields.
That's significant because many of those jobs were supposed to be the safe path into the middle class.
For years, the advice was simple: get educated, build skills, enter a knowledge profession.
Now AI is increasingly capable of doing the entry-level tasks that once helped young professionals acquire those skills.
Market research. First drafts. Basic coding. Data analysis. Routine legal work.
If AI can do the junior work, how do humans become seniors?
That's a problem bigger than job losses. It threatens the traditional career ladder itself.
The real danger isn't just unemployment
Clara Shih, a former technology executive at Salesforce and Meta, says AI agents can now allow a handful of people to accomplish work that previously required dozens.
She worries about more than paychecks.
Work gives people status, identity, purpose and community. Remove work at scale, and the consequences can extend far beyond the labor market.
The question isn't simply whether AI will take your job. It's what happens to society when millions of people feel their work—and therefore their value—is no longer needed.
That's where comparisons with globalization and factory automation become uncomfortable.
Technology can create enormous wealth while simultaneously leaving particular groups of workers behind.
The fact that society eventually adjusts doesn't make the transition painless.
So what should we do?
The answer isn't to stop AI.
Technology isn't going backward.
But we also shouldn't pretend that disruption automatically produces a better future for everyone.
Companies can choose to deploy AI as a tool that augments workers rather than eliminates them. Governments can rethink tax incentives and labor policy so that employing people isn't systematically disadvantaged relative to automation.
And workers need something more useful than vague advice to “learn AI.”
They need opportunities to develop skills that complement increasingly capable machines: judgment, leadership, creativity, domain expertise, communication, relationship-building and the ability to take responsibility when the answer isn't obvious.
Because the biggest mistake would be assuming the future of work is simply humans versus machines.
The real battle is over something more consequential:
Who gets to decide what AI is for?
If we use it primarily to eliminate labor, we may create extraordinary productivity—and extraordinary inequality.
If we use it to make people more capable, we could create something much better: a world where technology takes the drudgery out of work without taking the meaning out of it.
The future isn't predetermined.
But the window to shape it is closing.