Harvard College Dean David J. Deming encouraged students on Monday to learn how to use artificial intelligence, telling an economics class that the technology has spread faster than the personal computer or the internet and is already reshaping the labor market they will soon enter.
Deming, a labor economist who taught at the Harvard Kennedy School before becoming dean of the College last July, walked through several of his recent papers on AI adoption, its effect on jobs, and what it means for students during a guest appearance in Economics 50: “Using Big Data to Solve Economic and Social Problems.” He was introduced by Ec 50 instructor Raj Chetty ’00, a longtime collaborator.
Deming told students they needed to understand how to use AI because it was “going to be with us forever,” and urged them to experiment with the technology to figure out where it could help them.
“What is it good at? What is it less good at for me?” he said. “Think about it as a tool for, broadly speaking — not even self improvement, because that sounds very new agey — but it’s like your own personal assistant, and it never gets tired and doesn’t get mad at you for asking it questions.”
At the same time, Deming emphasized the value of learning without relying on AI and the importance of building skills “the hard way.”
“You can ask AI about anything, and it’ll give you the answer, but the skills you acquire in the process of developing expertise are more valuable than ever,” he said.
He told students to find a subject they cared enough about to want to learn on their own.
“You have to find something that’s intrinsically interesting to you, or it’s like, ‘I could get the answer from AI, but I don’t want to, I want to discover it on my own, and I want to learn it my own way,’” he said.
Deming overviewed several of his recent papers during the lecture, including one that found AI adoption has outpaced both the personal computer and the internet at equivalent points after each technology’s launch.
Advertisement
He also discussed what he called a “back of the envelope” finding that every hour spent using AI could boost a worker’s productivity by roughly a third.
“I know that it’s saving me a lot of time,” he said.
But Deming said those gains will not translate into broader economic growth until employers raise their expectations of what AI can do. Companies that fail to meaningfully incorporate the technology into their workflows, he added, would likely face layoffs and uncertainty in the process of “dying a painful death.”
“I think you will see a lot of that churn,” Deming said. “If AI ends up being a big deal, that’s exactly what the way it works, the creative destruction process.”
Still, he cautioned against treating the current moment as uniquely turbulent, arguing that AI-driven disruption is no greater than the large-scale transformations brought on by railroads, electricity, or the internet.
“Basically, we’re not in a unique period of disruption,” he said.
Deming also pushed back on the view that AI is already eliminating entry-level white-collar jobs, saying the data does not support that fear.
“There’s no relationship at all between the amounts of exposure your job has to AI and the speed of changing,” he said. “This doesn’t get covered in the news, but I will tell you, the best evidence is that there is no evidence that AI is particularly harming younger workers.”
Deming became a full professor at the Graduate School of Education and Kennedy School in 2016 and stepped back from the classroom to become the Kennedy School’s academic dean in 2021. Asked after the lecture whether he would return to teaching, he said he was not certain, but added that he hoped to during his tenure as College dean.
“I hope to teach something related to AI,” he said.
Advertisement
—Staff writer Sebastian B. Connolly can be reached at [email protected] or on Signal at sbc.23. Follow him on X @SebastianC4784.
—Staff writer Summer E. Rose can be reached at [email protected] or on Signal at ser.85. Follow her on X @summerellenrose.