“If AI is going to destroy all the jobs, why don’t we just stop?”

That was the rhetorical question my college-age son asked after we talked about the possibility of drastic changes to career paths and society thanks to AI (technically, generative AI). It was in line with what I’ve been worrying about myself.

Nobody really knows how disruptive AI will be. But young people and their parents would be foolish not to prepare for deep, unprecedented change in how we work. A huge portion of entry-level white-collar jobs — the kind that college graduates normally flock to and count on as career springboards — may not exist in the near future.

I’m not alone in these estimations, obviously. Dario Amodei, the CEO of Anthropic, has been brutally honest about what he believes his products will do to hiring. He has (repeatedly) said that half of entry-level jobs — especially in fields like finance, consulting, law, and tech — are likely to disappear within a few years. Interestingly, he’s changed his tune very recently, suggesting that there’s an opportunity for job growth. But either way, the facts on the ground bear out the concerns. Reductions have begun: Goldman Sachs estimates that 16,000 jobs are evaporating every month.

So, what’s to be done? In a widely circulated clip from a May 2024 interview, former Google CEO Eric Schmidt put it plainly: Once AI agents develop a suite of skills that allow them to start working together on their own, away from our guidance, “we won’t understand what the models are doing.” His suggested solution? “Pull the plug.”

It’s a gut response that I feel a strong affinity toward, even as I dive deep into using AI myself. As I watch the world barrel toward a truly unknown future and the potential devastation that AI could wreak on job markets and young workers, I feel a mounting unease about how companies are starting to respond.

What makes this challenge particularly hard to solve is that the executives making decisions about AI deployment and jobs will be fine regardless of how this plays out. They have capital, seniority, and options — financial and otherwise. It’s sadly uncommon for leaders to think beyond market cap and their own vesting schedules and consider whether we all can thrive. That inequality in exposure to risk is part of what makes this more than just a business question.

When society faces deep risk, companies and leaders tend to make choices that seem optimal for their short-term interests. From a pure short-term-profit perspective, bringing in fewer workers is probably the financially smart thing to do. But thinking about only the short term poses significant danger. With this latest existential challenge, if companies continue to head down a “people-light, AI-token-heavy” path, the risks aren’t just to young workers but to businesses, too.

The microeconomic case for some degree of caution is this: If companies decimate entry-level roles, what happens to the pipeline for leadership? Service businesses have long had a pyramid model where lots of young, smart kids come in and get trained and tested, and then a small subset make it to partner or other senior roles.

So, what if companies just didn’t eliminate as many jobs? Yes, we’re about 40 years into this model of businesses announcing cuts and their stock rising — investors often love companies that fire people. But what if, this time, they just didn’t? The companies that preserve human judgment, build institutional knowledge, and keep developing talent may find themselves with the advantage down the road.

* * *

We know AI isn’t going away. What it can already do can feel like magic. And its use will rise as companies mandate it and people discover what it does well, acting as their assistant, researcher, editor, and more. But GenAI has some serious issues and flaws, such as its tendency to hallucinate. And its footprint and effect on communities is enormous. I write this as a practitioner watching this unfold up close, and as someone who uses AI every day and is actively working on how to reduce its energy footprint.

But I have a sneaking suspicion that we will look back at early 2026 and kind of wish we had just stopped. Of course, this won’t happen writ large. There is global geopolitical competition, and there are stunning amounts of money to be made.

We have the option to make wise, thoughtful choices about how we treat employees — you know, the people who actually make up a thriving economy by having jobs and disposable incomes to buy things. The leaders with the power to make the call about how people are cared for will land on their feet either way. The ones just entering the workforce — my son’s generation — may not have that luxury.