Shorting the AI jobs apocalypse
Turbulent times are windows of opportunity for well-educated and ambitious young people
Hi everyone-
Thanks for sticking with Forked Lightning! I’ve been quiet for a little while, mostly because I took a new job as Dean of Harvard College. This leaves me with less time for research and writing than I’d like. Still, I’ve been doing my best to keep up with the latest research on the labor market impacts of AI.
I’ve got two things on the menu today. First, I wanted to react to a new paper about AI and entry-level jobs that’s caught everyone’s attention. Second, I gave a speech to the Harvard Class of 2029 yesterday about why an AI economy could be a huge win for them and for other young people. I’ve reproduced the speech below – read on if you’re interested!
Canaries in the Coal Mine
Like everyone else I read with great interest the recent paper by Brynjolfsson, Chandar, and Chen, “Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence.” This paper uses administrative data from ADP (a large payroll company) to track employee headcounts by age and occupation before and after November 2022, when ChatGPT was introduced to the world. It shows a bunch of charts like this:
The authors emphasize the headline finding, which is that employment for young workers (age 22-25 in blue, and age 26-30 in orange) has decreased relative to its pre-ChatGPT trend and also relative to older workers in the same occupation. The figure above shows software developers, which is the starkest result, but they also find a similar pattern for other “AI-exposed” occupations like customer service representatives.
Apocalypse….Now?
Many others have written about this paper, so I’ll just make three quick points. First, while the figure above shows that young workers are suffering more, it’s a little misleading. Those lines each represent a group of workers, but they are NOT the same size. Your eyes are drawn to the blue line, but it only has 4 years in it. The top line (in purple) has 9 years in it (ages 41-49). If age were uniformly distributed, those two lines would roughly cancel out when weighted by size. If you smashed all the lines together, you’d see that software developer jobs have continued to grow overall through the release of ChatGPT.
Second, the paper’s story is about declining relative labor demand for young workers. Usually when the demand curve shifts, price and quantity move in the same direction. In this case, we’d expect wages (e.g. the price of labor) to fall when employment is declining. But in fact, they show the opposite. The top left panel of Figure 10 shows compensation changes by age group, and young software developers seem to be earning more, not less.
Third, while the paper is very careful about maintaining a consistent sample of firms, this shuts down a potentially important margin of adjustment. They restrict the sample to firms that have continuously employed workers from 2021 to 2025. But we know that AI adoption is faster in smaller firms, including startups. We also know that big, incumbent firms are slower to adopt new technologies. Even if entry-level employment really has declined in big legacy firms because of AI, it might still benefit young workers because they are forced to become entrepreneurs or join startups. The employment decline on the graph could reflect a shift out of incumbents who use payroll services like ADP, and into new competitor firms that aren’t captured in the data. There has been a surge in new business formation since the pandemic, so sticking to big established companies might miss an important part of the employment picture.
Despite everything, I’m still very bullish on the future of entry-level college graduates in an AI-fueled economy. My speech yesterday at Harvard Convocation for the Class of 2029 makes this argument in much more detail.
See below!
*************************************************************************************************************
Hello, Class of 2029!
On behalf of the faculty and staff, it is my sincere honor to welcome you to Cambridge, welcome you to Harvard College, and welcome you to what I hope and expect will be the journey of a lifetime.
We share a very special bond. This is a first year for both of us— you as students and me as your Dean. We’re on this journey - this thrill ride - together. Let’s look out for one another. We’ll share our triumphs, and our challenges. And, if you have first day jitters, let me assure you, you’re not the only one.
You come to college at a unique moment in history. There’s a lot happening in the world – and, let’s be honest – there’s a flashing, bright spotlight on Harvard. Many people have questions about the promise of an education like this, and yet despite all that, we’re here, because we believe in this place. And we believe in the power of a Harvard education and its ability to lift people up, to serve the nation, to change the course of history, and to improve the lives of others. You, and the classmates sitting around you, wanted to be right here at Harvard in this moment. You’ve chosen to be in the eye of the storm. You’ve chosen to be a part of history.
I think that’s really admirable. Although I must confess - my belief in your valor is selfishly motivated. Like you, I accepted this job recently, knowing exactly what I was getting into.
Why did I do it? (many, many people have asked…) It’s simple. Harvard matters in the world, and it’s a great gift to be given the opportunity to contribute to something that truly matters.
It is my fondest hope for you, Class of 2029, that you will learn and grow in wisdom and knowledge. And more importantly, I hope that when you leave and you are choosing what to do with your Harvard education, you’ll be skeptical of safe, established, well-worn paths, and instead do what you are doing today - leaping boldly into the unknown. I believe that you will be a class of creators, builders, and dreamers.
We are living in a time of uncertainty and fear about the future— fears that extend far beyond these gates. There’s a broader sense of foreboding in the air about the value of education in an age of Artificial Intelligence. As a labor economist, it’s something that I think and write about a lot, and something I know has already affected your lives directly. It’s also not lost on me that we are gathered on Labor Day, when we pause to celebrate the American worker. You have probably heard that some of the large white-collar employers that pay nice starting salaries and typically snap up Harvard graduates by the dozens each year are suddenly pulling back, often citing AI as the reason for their caution.
Still, I’d go long on the value of your Harvard education. I predict that an AI-infused economy will make a liberal arts and sciences education more valuable, not less, because one of the most important skills you learn here is how to adapt to a changing world. And change is unavoidable.
CEOs really are talking a lot about AI, but that’s partly because it’s a convenient scapegoat. If your company’s suddenly struggling and you are looking around for someone to blame, AI makes a great target. Chatbots don’t fight back. And if you have to explain to shareholders why you missed your earnings target, you’ll point the finger at the deus ex machina of technological apocalypse, rather than at your own decision-making.
Having said all that, I do sympathize with those who fear for the future of entry-level college jobs at big companies. AI really can do the work of new college graduates – writing, research support, analyzing and synthesizing data – better than it can do the work of say, a plumber or a gardener.
AI makes formerly safe career ladders seem wobbly. As students, you’ve suddenly been given access to a technology that could write your essays and answer the hardest questions on your PSets in minutes (which you really shouldn’t do). It’s fair to wonder whether this weird, alien intelligence that’s available at the low price of free or $20 per month has reduced the need in the broader world for young, talented Harvard students like yourselves.
But I want to be clear, I believe this doom and gloom is misplaced. (If I thought otherwise, I wouldn’t be bringing it up on a joyous occasion like this.) Traditional, “safe” pathways in big established companies may very well be at risk from AI. But you, Class of 2029, have lots of other options. You don’t need to follow those well-worn paths. In fact, you may be better off without them. History tells us that periods of technological disruption and economic distress are windows of opportunity for educated, ambitious young people like yourselves to join – or start - something new.
It’s no accident that many of today’s top companies were founded during recessions or in the “quantum foam” of past breakthroughs like electricity, computers, and the internet. AI – like earlier technological revolutions – may require what our colleague Rebecca Henderson calls “architectural” innovation”, fundamentally rebuilding production and design from the ground up. This is harder for big companies – like those that hire scores of Harvard grads every year - because they are physically invested in old ways of doing things.
Disruptive innovation pushes out incumbents and creates new winners and losers in the economy. We may be living through such a moment right now. If that’s true, you don’t want to fight the wave. You want to ride it, while it’s cresting. Do we have any surfers in the audience? Well, you know that when it comes to riding big waves, timing is everything. In that sense this moment, which feels risky, is actually a tremendous opportunity for you.
Looking back, if you graduated from Harvard during the mid 1990s, when the internet was really taking off, the safe jobs would have been at corporate giants like ExxonMobil, Coca-Cola, Bank of America, and Procter and Gamble. The uncertain path would have been a startup like Google or Nvidia. As a matter of fact, do you know how many of the top 10 companies in 1990 by market capitalization are still in the top 10 today? You can count them on no hands. Zero.
My point is that your career is going to last a long time - 40 years or longer. While there’s a decent chance your “dream job” at a steady company will one day be downsized or disrupted, there’s also a good chance that some of the founders of top 10 companies in 2050 are sitting in the audience right now.
Not everyone wants to join a startup or become an entrepreneur, nor should they. But the opportunities created by disruption and rebirth extend far beyond the private sector. Sal Khan left his job at a hedge fund to start Khan Academy in 2008, during the job wreckage of the Great Recession. Today, millions of students treat it as a second teacher. Around the same time, a group of Harvard and MIT students launched GiveDirectly, which uses mobile payment technology to deliver cash to poor households around the world at extremely low cost. New technologies can create exciting career opportunities, but they can also be deployed to benefit humanity.
Over the last century, disruptive innovation has generally favored the young and the well-educated, because they can adapt to new ways of doing things. AI could be different, but I doubt it. Young educated people like you are already the heaviest users of AI, and you are creative and open-minded enough to figure out the best uses of it too.
And – if I may be immodest and talk up our own work - this is where the liberal arts and sciences education we deliver here at Harvard really shines. It’s a general purpose, future-proof toolkit. You’ll learn here how to think critically and how to understand others who are different from you. You’ll also figure out the kind of person you want to be, what truly matters to you, and what you want to contribute to the world. These are timeless lessons. Career education is important too, but it can have a shelf life. Coding bootcamps were all the rage a few years ago, and now people are using AI to “vibe code” in conversation English. Things change, faster than you expect.
We are likely living in an interregnum between the pre- and post-AI economy. I don’t know what’s going to happen, but I do think AI is going to shake up the snow globe.
So my advice is to be a creator, a builder, and a dreamer. Take intellectual risks in the classroom and in your own career choices, not because risk is inherently a good thing, but because adherence to the status quo is its own kind of risk, especially in turbulent times.
I believe that you, Class of 2029, by choosing (despite everything) to come to Harvard at this momentous time in our university’s – and our nation’s – history, will be the future leaders that will help all of us geezers up here in regalia figure out what’s next. Embrace the future. Be fearless and be interesting. Dream big dreams. Let’s leap boldly into the unknown, together.
Welcome to Harvard, Class of 2029!
Should we be down playing the impact of AI on young people, youth unemployment and the demand for entry level jobs? To be honest youth unemployment is a problem in many regions and countries and hiring freezes are a sign of a very restricted economy. Well no more I'm assuming with the jobs report later today.
If millions of students are using chatGPT can we assume that even the value of a college education will become eroded? Are we supposed to imagine that these young people are more likely to become entrepreneurs because of their cohort struggle to find an entry-level job?
If the Trump Administration wants to tweak immigration and deport masses of people, it's going to have dire consequences on the labor demand supply market. Meddling in industrial policy is not helping. If tech companies continue to use AI for a greater amount of their code, what is that even mean for the long-term demand for software engineers?
That’s a great speech David! It got me excited about the future too.