The (Job) Sky Is Falling. No, Seriously
When I came in as head of NC State’s Institute for Emerging Issues in 2017, one of the first projects I took on was trying to get people to pay attention to a report we had published called “FutureWork.” The report bundled a bold and dire set of data and projections. Automation was accelerating. That was going to mean great increases in efficiency for the economy, but it also meant that, over time, there would be huge job loss – our projections showed by 2040, there was a 70% chance that 25% of our current jobs would be automated out of existence, with numbers as high as 41% in some counties of North Carolina.
I shared the results with state agency heads and educators and employers across the state, arguing we needed to make dramatic changes to how we educated people and to start getting folks ready to reinvent themselves every few years. We needed to reimagine our unemployment and workforce training systems. People nodded along for about five minutes, then yawned. “It’s not going to be as bad as you think. And it’s not going to be as soon as you think. I’ve got bigger problems to deal with right now. Why don’t you go back to that think tank and think about how to help me with REAL problems.”
The gaslighting worked: I was new in my job and these were smart people. Maybe I was Chicken Little. I moved on to other emerging issues.
Eight years later, it’s clear that the folks I talked to about job loss were right on at least a couple of things – it wasn’t going to be as bad as I thought or happen as soon as I thought – it was going to be worse and happen sooner.
It’s an AI thing.
These days, it’s Dario Amodei, much smarter, with much more data and credibility in hand, who’s sounding the jobs alarm. I really hope more people will pay attention to him.
Amodei doesn’t plan to stop making AI; he does want more policymakers to begin planning for its implications.(Photo Benjamin Girette, Bloomberg)
In a recent interview, Amodei, the 42-year-old CEO of Anthropic, one of the country’s top producers of artificial intelligence, told Axios it was time for AI execs to stop “sugarcoating” what he believes to be the truth: that AI could wipe out half of all entry-level white collar jobs, particularly in technology, finance, law and consulting, and could boost overall unemployment to 10-20% in the next one to five years.
The job market is already changing, as companies begin to move beyond using AI to augment people and start getting ready to use it to replace people.
· A new study by Oxford Economics finds that intern hiring by companies is down and “there are signs that entry-level positions are being displaced by AI at higher rates.” Handshake reports that more than half of current college seniors are pessimistic about the job market, with the average graduate applying to 21% more jobs this year than last. Unemployment for 22-27 year olds at 6% compared to 4% for all workers. The New York Fed has observed that the job market for recent college grads has “deteriorated noticeably.”
A Glassdoor survey finds employees more worried about the future of their jobs than at any point over the past decade.
· Already-hired entry-level workers aren’t safe either. LinkedIn’s “chief economic opportunity officer,” Aneesh Raman noted that AI is breaking “the bottom rungs of the career ladder – junior software developers…junior paralegals and first-year law firm associates” as well as young retail associates and customer service reps.
· And it’s not just entry-level workers. Mark Zuckerberg told Joe Rogan in January that by the end of this year “we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of mid-level engineer that you have at our company that can write code.” Not long after that Meta made an announcement: it’s cutting its workforce by 5%.
· Microsoft is cutting about 3% of its workforce, many of them engineers.
· Wal-Mart has cut 1500 corporate jobs in anticipation of the shift ahead.
· Crowdstrike, a cybersecurity firm, is cutting 5% of its workforce, noting we are at “a market and technology inflection point, with AI reshaping every industry.”
So far the big changes are coming in technical areas, but as technology spreads, it won’t stop there. A San Francisco company called Mechanize isn’t the least bit shy about sharing its ambition: “We’ll only truly know we’ve succeeded once we’ve created A.I. systems capable of taking on nearly every responsibility a human could carry out at a computer, the company wrote in a recent blog post. “Our goal is to automate all work,” says 29-year-old Tamay Besiroglu, one of the company’s co-founders, told the New York Times.
“We want to get to a fully-automated economy, and make that happen as fast as possible.”
Besiroglu and his other co-founders disagree only on how fast that is -- 10, 20 or 30 years?
OK big deep breath. Let’s admit it is part of the culture of AI co-founders seeking huge amounts of cash and computing power is to sell big stories, to “fake it till you make it,” to overpromise, then underdeliver. So let’s say that there’s only a 50% chance that Amodei is right about half of entry level jobs going away in 1-5 years and there’s only a 10% chance that Mechanize is right that all jobs involving a computer will be automated in 10-30 years. Either way, it’s completely irresponsible for us not to be having big, ugly, ongoing, thoughtful discussions about what the future of work will look like.
So who’s thinking about the implications of the job change that is coming?
Not the Trump administration: they’re focused on making sure the US wins the AI race with China, and the “Big Beautiful Bill” in its current form actually prohibits any state from putting any restrictions on AI. Their perspective: we need to win first, figure out how to deal with the consequences later. But Trump’s own supporters warn that they ignore the jobs issue at their own peril: as Steve Bannon noted recently on his “War Room” podcast, “We have to get ahead of this or we’re going to have mass unemployment, particularly among entry-level people under 30.” MAGA activist Charlie Kirk echoed that idea: “What you are going to see is one of the most dramatic job displacements, and it’s going to be a top issue in the 2028 campaign.” Mass unemployment will be politically inconvenient.
Not the folks at Mechanize or most AI companies. They’re focused on building ever-smarter, ever-more-powerful AI that enables them to hit their goals of a fully-automated workplace. In an interview with the New York Times, co-founder Matthew Barnett made clear why he thinks the tradeoff was worth it.
“If society as a whole becomes much wealthier, then I think that just outweighs the downsides of people losing their jobs.” Mr. Barnett said.
Does a wealthier economy actually outweigh “the downsides of people losing their jobs”? (Image generated by ChatGPT)
But Amodei is thinking, and worrying about it (while still releasing a string of new AI products): “We, as the producers of this technology, have a duty and an obligation to be honest about what is coming… We’re saying: ‘You should be worried about where the technology we’re building is going.’ Critics reply: ‘We don’t believe you. You’re just hyping it up.’”
Instead, he says, the critics should ask themselves, “’Well, what if they’re right?’”
Anthropic has created a couple of useful vehicles to help think about job loss. Their Anthropic Economic Index attempts to track data about how AI is being deployed in various fields, including software development and the Anthropic Economic Advisory Council aims to provide thoughtful commentary on the economic and labor market implications of AI’s growing power. And Amodei has even floated the idea of a “token tax” on AI companies: every time someone uses an AI model and the AI company makes money, maybe the government should college a 3% tax to redistribute to those harmed by AI.
It makes sense that an AI leader is sounding the alarm – who knows better where the technology is going? – but AI’s interests will never be the same as the broader interests of society. Government leaders, educators, private sector employers, nonprofits and real people need to start learning and thinking and talking and working with AI (or without) to shape what this new world looks like.
Seriously, is there a priority bigger than to start figuring out what humanity is going to DO for the rest of our existence? Let’s get on that now.
It’ll be scary and hard and most of the options are more likely to be “least worst” rather than “best,” but that’s no excuse for putting our heads in the sand. “You can’t just step in front of the train and stop it,” notes Amodei. The only move that’s going to work is steering the train – steer it 10 degrees in a different direction from where it was going. That can be done. That’s possible, but we have to do it now.”
Of course, starting eight years ago would’ve been even better.
Notes:
Oxford Economics study on changing job market: https://www.oxfordeconomics.com/resource/educated-but-unemployed-a-rising-reality-for-us-college-grads/
Amodei interview with Axios: https://www.axios.com/2025/05/28/ai-jobs-white-collar-unemployment-anthropic
Latest Glassdoor Employee Confidence Survey: https://www.glassdoor.com/blog/glassdoor-employee-confidence-index-may-2025/
Challenges for recent college graduates: https://www.nytimes.com/2025/05/30/technology/ai-jobs-college-graduates.html
Mechanize projections of full automation: https://www.nytimes.com/2025/06/11/technology/ai-mechanize-jobs.html?smid=nytcore-ios-share&referringSource=articleShare
Big Beautiful Bill provisions restricting AI regulation: https://www.theverge.com/ai-artificial-intelligence/684924/congress-big-beautiful-bill-state-ai-law-ban-pushback
Anthropic Economic Advisory Council: https://www.anthropic.com/news/introducing-the-anthropic-economic-advisory-council
Anthropic Economic Index: https://www.anthropic.com/research/impact-software-development