Let's be honest, it’s hard not to feel a little anxious about AI and our jobs.
Scroll through your feed, and you’ll see headlines that feel like they’re straight out of a sci-fi movie. You’ve got CEOs of major AI companies talking about a future where AI could do all human jobs in less than five years. One researcher even predicted a near-term recession and a "breakdown of the early-career ladder."
It’s enough to make you want to unplug everything and go live in a cabin. And who could blame you? When the people building the technology sound this grim, the panic feels justified.
But what if I told you we might be panicking about the wrong thing? What if the way we’re measuring AI’s threat to our jobs is completely missing the point? I had a chat with an economist who thinks so, and what he said completely changed how I see this whole situation.
Why Our Current Guesses About AI and Jobs Are Basically Useless
Right now, when you hear that a job is, say, "28% exposed to AI," what does that even mean?
It usually comes from studies by big names like OpenAI or Anthropic. They look at a massive government catalog of all the individual tasks that make up a job. Think of it like a giant recipe book for every career. Then, they check off which of those tasks an AI could theoretically do.
For example, part of a real estate agent's job is asking clients what they're looking for. An AI can do that. Check. So, that task is "exposed."
Here's the problem, and it's a big one: knowing that an AI can do a task tells you almost nothing about whether it will actually replace a human.
Alex Imas, an economist at the University of Chicago, put it bluntly when we spoke. He said that looking at "exposure alone is a completely meaningless tool for predicting displacement."
Think of it this way: a microwave can cook a meal, but that doesn't mean every home cook is at risk of being replaced by a machine. You still need a person to decide what to eat, prep the ingredients, and know not to put a fork in there. For most jobs, it’s the same story. AI might handle a few tasks, but not the whole role.
The Real Question We Should Be Asking
So, if "exposure" is the wrong metric, what's the right one? It all boils down to a single, crucial question that Imas says should be keeping policymakers up at night.
Let's imagine a software developer who builds dating apps. With new AI coding tools, she can now do in one day what used to take her three. She's suddenly way more productive.
Her boss is thrilled. The company is getting more output for the same amount of money. But what happens next? This is the fork in the road.
- Path A: Growth. Because the company can build apps faster and cheaper, they lower their prices. Suddenly, millions of new users sign up. The demand explodes. To keep up, the company has to go on a hiring spree, bringing on more developers.
- Path B: Layoffs. The company lowers its prices, but demand barely budges. It turns out, the people who weren't using premium dating apps still don't want them, even if they're a bit cheaper. Now, the company needs fewer developers to do the same amount of work, and layoffs begin.
Which path do we go down? The answer to that question, repeated across every single industry, will determine the future of work. And right now, we are flying completely blind.
It All Comes Down to a Supermarket Secret
So, how can we possibly predict which path a company or an industry will take? The answer, surprisingly, has a lot to do with a box of cereal.
Imas explained that the key is a concept called price elasticity.
It sounds complicated, but it's actually simple. It's just a measure of how much demand for something changes when its price changes.
We have tons of this data for groceries. The University of Chicago partners with supermarkets to get data from their price scanners. They know exactly how many more boxes of Cheerios people buy when the price drops by 50 cents.
But here’s the scary part: we have almost none of this data for jobs and services.
We have no idea how much demand for a personal tutor would increase if AI made their services 30% cheaper. We don't know if people would hire more web developers if their rates dropped. We can't predict what would happen with dietitians, writers, or project managers.
It’s like trying to navigate a ship in a storm with no map and a broken compass.
The "Manhattan Project" for Our Economic Future
This is where Imas gets serious. He says we need something like a "Manhattan Project to collect this" data.
We need a massive, coordinated effort to start tracking these numbers across the entire economy. Not just for the jobs that seem obviously affected by AI today, but for everything. Because, as he points out, "Fields that are not exposed now will become exposed in the future."
Yes, it would take a lot of time and money. But the alternative is to keep guessing, keep panicking, and let our policymakers stumble in the dark while our livelihoods hang in the balance.
Getting this data would give us the first realistic glimpse into our AI-enabled future. It would allow us to see which industries are likely to grow and which are headed for trouble. It would give our leaders a fighting chance to create actual, effective plans—like targeted retraining programs or economic support—instead of just talking in vague platitudes.
So, the next time you see a headline about AI coming for your job, take a breath. The real story isn't about whether a robot can do a task you do. It's about whether AI will make your work so much more valuable that the world demands more of it. And it's time we started demanding the data to find out.
