The AI Is Taking Our Jobs!

The Atlantic’s Derek Thompson points to “A New Sign That AI Is Competing With College Grads.”

Something strange, and potentially alarming, is happening to the job market for young, educated workers.

According to the New York Federal Reserve, labor conditions for recent college graduates have “deteriorated noticeably” in the past few months, and the unemployment rate now stands at an unusually high 5.8 percent. Even newly minted M.B.A.s from elite programs are struggling to find work. Meanwhile, law-school applications are surging—an ominous echo of when young people used graduate school to bunker down during the great financial crisis.

He acknowledges that the causes are likely multiple, including the restructuring of the economy in the wake of the Great Recession and a gradual decline in the wage benefits of college. But he focuses on the titular explanation.

The third theory is that the relatively weak labor market for college grads could be an early sign that artificial intelligence is starting to transform the economy.

“When you think from first principles about what generative AI can do, and what jobs it can replace, it’s the kind of things that young college grads have done” in white-collar firms, [David Deming, an economist at Harvard] told me. “They read and synthesize information and data. They produce reports and presentations.”

[…]

As law firms leaned on AI for more paralegal work, and consulting firms realized that five 22-year-olds with ChatGPT could do the work of 20 recent grads, and tech firms turned over their software programming to a handful of superstars working with AI co-pilots, the entry level of America’s white-collar economy would contract.

[…]

 And even if employers aren’t directly substituting AI for human workers, high spending on AI infrastructure may be crowding out spending on new hires.

And even if employers aren’t directly substituting AI for human workers, high spending on AI infrastructure may be crowding out spending on new hires.

This is all admittedly quite thin: a trend analysis of something that’s barely a trend. It is, however, quite plausible.

When ChatGPT exploded onto the scene two years ago, it sparked a lot of angst in higher education circles, with administrators and professors alike seeing it as a vector for cheating and panicking to figure out how to make it harder to do. Since then, most of us have come to embrace it as yet another tool available to students and professors alike and have redesigned both teaching and assessment around the fact that it exists.

I haven’t spent as much time as I should experimenting with it yet, as early attempts weren’t that promising. When I was department head, I found it somewhat helpful at helping me come up with assignment prompts that were better than I would have come up with on my own. And my wife has been using it a lot a work to automate mundane writing tasks.

As a colleague who’s an enthusiastic adopter keeps noting, today is as bad as the technology will ever be. He’s already using it to help write op-eds, PowerPoint presentations, and even generate podcasts.

I suspect that I’d go back to being a more prolific blogger if I fed interresting pieces into AI and let it do the tedious task of linking, summarizing, finding appropriate graphics, and the like, reserving only the analysis for myself. (Andrew Sullivan (in)famously did that with interns many years ago now.) Getting to that point, though, requires investment of significant up front time getting proficient with the tools.

But, getting back to Thompson’s point, this also means that an increasing number of tasks that used to require highly intelligent, well-trained people to accomplish will be able to be essentially automated. Aside from the economic displacement and social upheaval that will cause—how will these people make a living or find meaning in their daily existence?—there’s also the matter of the tools being smarter than the operator.

We’re already at the point where the best AI programs are better than the average undergraduate at writing a 2000-word, thesis-driven essay based on an assigned body of literature. A year or two from now, they’ll likely be better than the average PhD student. Certainly, they’ll be massively faster and more productive.

But who’ll be in charge of checking the work? As technology has taken over tasks that we used to do with human brainpower, including simple information storage, the incentive to actually know things has diminished greatly. (There was a time, for instance, when I had dozens of phone numbers and addresses memorized.) We can, after all, just Google it.

Having bridged those worlds, I have the advantage that I have a reasonable ability to separate wheat from chaff in the search results. But, as we offload more and more of our cognitive tasks to machines who are much better at storing and processing information, who is going to have that skillset?