Will AI Eventually Replace Human Workers or Augment Them?

Will artificial intelligence (AI) make human workers obsolete — or make them more valuable than ever? That is the burning question in the age of AI, one that understandably concerns many workers.

But MIT economics professor David Autor argues that AI will end up generally augmenting workers instead of replacing them.

“There are two competing visions of AI. One is machines make us irrelevant. Another is machines make us more useful. I think the latter has a lot to recommend it,” said Autor, during a keynote address Tuesday (March 31) at the 2025 MIT AI Conference in Cambridge, Massachusetts.

He cited a historical precedent. “Over the last 200 years, we have automated so much of what we do. We have mechanized. We have moved ourselves out of agriculture, out of manufacturing, out of back-breaking toil,” Autor said. Yet, “We have made labor more valuable during that period.”

Not only have wages risen, but labor’s share of all economic activity has gone up as well and been stable, at 60% of GDP.

What made labor valuable was expertise, which Autor defines as domain-specific knowledge that lets a worker accomplish a bigger goal. The skill has to have economic value and it can’t be something just anyone can do.

“If everyone is an expert, then no one is an expert,” Autor said.

However, workers remain fearful. A January 2025 PYMNTS Intelligence report, “GenAI: A Generational Look at AI Usage and Attitudes,” shows that 54% of survey respondents think AI poses a “significant risk” of widespread job displacement. These fears run across industries and demographics.

But some are more worried than others: Those who work in technology and non-customer-facing roles are the most concerned (58%), while healthcare and education workers are less so, at 48% and 52%, respectively.

Baby boomers and Gen X workers are more worried than Gen Z. High-income workers with college degrees have “above-average” levels of concern. The report is based on a survey of nearly 3,000 U.S. consumers conducted in November 2024.

Read more: IMF Says AI Could Impact 60% of Jobs in Advanced Economies

To be sure, AI is eroding parts of human expertise. By learning patterns from massive datasets, it is performing tasks once thought to be uniquely human, such as writing poems or making subjective analyses.

Autor cited the example of CheXpert, an AI system that reads chest X-rays and outperforms many radiologists in diagnostic accuracy. This type of work is suited to AI because there are no hard and fast rules — like a suspicious shadow in an X-ray must be a certain size, for example. Instead, the AI learns to spot suspicious anomalies by learning from troves of patient data.

But automation, Autor warned, is not full proof. In the case of CheXpert, radiologists working with the tool actually made more errors than those working alone. Why? Because they didn’t know when to trust the machine — and the machine wasn’t built to help them learn that judgment.

With these mixed results, the future of labor in the age of AI will depend on what people will do with it.

“The future is not a forecasting problem. It is design problem, something that we are collectively building,” Autor said. “There are many branches and choices to take about how we want to use this technology.”

Read more: New AI Agents Take on Management Jobs

Autor showed an image of mining workers in Congo manually hauling sacks. They did not have any axes, picks, helmets, respirators or other equipment that are readily available today.

Better yet, the mining company could have used machines to do the work. Why wasn’t mining equipment used?

“Because labor is so cheap in so much of the world,” he said.

So even with the adoption of AI, if the economics of labor work out better, human workers will still be used.

Second, the industrialized world is facing a shortage of working-age adults. Countries like Japan, Korea, Greece and Poland are on track to lose up to 40% of their 20- to 64-year-old populations over the next 35 years, he said. AI can fill the gap, not replace workers.

Third, technology has been used throughout the centuries to enhance human skills. AI will be no different.

The invention of the stethoscope did not remove the need for a doctor, and the pneumatic hammer did not put the roofer out of work.

“Tools often augment the value of human expertise,” Autor said. “They shorten the distance between intention and result. They enable us to do things we could not otherwise do without them.”

New technologies can also create the need for new expertise, he said. While some jobs would go away — such as transcriptionists — new job categories would be created.

Finally, Autor pointed out that the collaborative nature of work makes humans indispensable.

For example, He said his Wi-Fi-enabled washing machine lets him start a load of laundry remotely through a smartphone app but still requires him to physically remove the clothes.

“My washing machine has more processing power than the Apollo Guidance Computer of 1966” that was on board NASA’s Apollo Lunar Module, Autor said. But the washing machine has “an utterly worthless technology. I still have to go into my laundry room to get my clothes out.”

Photo: David Autor, MIT economics professor | Credit: MIT livestream

The post Will AI Eventually Replace Human Workers or Augment Them? appeared first on PYMNTS.com.