I recently had the opportunity to speak with NPR's Scott Detrow about the changing landscape of work, college, and AI. These are questions many of us are navigating so I wanted to share my expanded thoughts here for readers who follow my ongoing analysis of how AI and technology are changing the future of work.
What We Got Wrong About AI and Jobs
Scott asked me what the biggest misconceptions have been when it comes to AI and the workforce. I think it's important to note that we will always get future predictions wrong. The business of predictions in and of itself is flawed. But the forecasts did initially show that more routine tasks would be first in line for automation.
We pictured factory automation and basic data processing. But thanks to advances in generative AI and large language models, we're seeing something much more complex: AI is now capable of handling aspects of creative, unpredictable tasks, too. For instance, writing in and of itself was never viewed as a statistical prediction problem. Now AI can write better than most people (it might quote nonsense and be painted with a bland brush, but the writing itself outshines most people). AI can draft fairly strong marketing proposals, it can look up case precedents and synthesize legal arguments, it's writing music.
Scott asked me for a specific example of jobs suddenly at risk as a result of these capabilities. But it's more nuanced than that. First, we are seeing companies slow hiring over sudden layoffs. And second, it's about the nature of tasks within jobs and what percentage of a role those tasks represent. Even within the same department, AI is going to have a different impact on employment.
For instance, AI can now handle basic writing, data entry, analysis, marketing copy, first drafts of legal briefs, social media captions, and even basic coding for simple apps. These are all tasks that AI can complete at a level that would match what most junior hires would produce. Yes, there are mistakes and hallucinations, but with proper oversight, companies are asking themselves: if AI can handle 60-80% of the tasks in a role, does it make sense to hire a human for that position at all? In the case of a junior lawyer, it may not. But the junior lawyers tasks don’t entirely disappear, they move up the ladder.
I was chatting with a friend who is a senior associate at a law firm. He said the work he would have asked a junior associate to do he now does himself in partnership with AI. "It doesn't make any sense for me to bill a client for the junior's hours when it now takes me 30 minutes to complete, which is only slightly longer than the time it would have taken me to explain everything I needed."
Rethinking Education for an AI Future
The conversation naturally turned to how people should be thinking about their education and what they need to learn to stay competitive.
There are certain non-negotiable skills for thriving in an AI-driven future:
Judgment has become profoundly challenging in new ways, and in many ways it will increasingly become a technical skill. For instance, understanding how to weigh trade-offs and odds both when it comes to knowing when to seek AI, and how to evaluate the output it gives you.
Deep thinking takes on new importance when you need to structure problems effectively before handing them to AI systems and think across discrete functions and disciplines. This isn't the kind of thinking we typically focus on in school, but it's becoming essential.
Communication isn't just about communicating your problems well to AI but also to teammates. Economic research shows that as cognitive tools like AI amplify our capabilities, the cost of coordination and communication within organizations becomes increasingly critical. Affecting everything from recognizing opportunities to making effective decisions across teams.
Continuous Learning The average half life of most technical skills now hovers around two years. Being prepared to go back to the drawing board, learn new things, and pivot into different industries is the new fabric of the workforce.
Advice for Career Planning in Uncertain Times
We zoomed out to chat about the bigger picture: how people should make decisions about majors, industries, and career paths when the future is so uncertain. My take here is that we need a fundamental mindset shift.
The old model of "learn, work, retire" is disappearing. Students need to see themselves as entrepreneurs who bring a bouquet of skills and the ability to continuously adapt and learn. Moving through many different industries and job types shouldn't be seen as a sign of instability.
If you can think deeply, learn quickly, communicate well, exercise strong judgment, and adapt to change, you have a fierce toolkit for the future of work. It doesn't really matter which way the technology goes or how industries shift, you can pivot and evolve with whatever comes next.
It's much easier to predict which jobs might be automated than which new ones will emerge, so I always come back to this: invest in the skills that help you learn and adapt, rather than betting everything on one particular career path.
That said, there are sectors showing clear growth potential. Biology represents a major frontier where tech and AI are converging. Space exploration and robotics offer similar opportunities. But I think there still has to be an underlying interest in the fields. I wouldn't recommend someone with a disdain for robotics simply chalk that up and march into a four-year degree for the sake of moderate (but still unpredictable) certainty.
You can read my analysis on the End of the Career Ladder and my advice for college students here.
The Question I Ask Myself Daily
Scott's final question: "How often are you actively thinking about how different your job will be in five or 10 years?" My answer: all the time.
Here's what I know that gives me both urgency and comfort: AI is a general-purpose technology. Unlike other innovations that come and go with hype cycles and bubble bursts, general-purpose technologies become infrastructure. We rebuild our societies on top of them. That much is guaranteed.
This transformation is coming whether it happens in three years, eight years, or fifteen years. The timelines might be fuzzy, but the direction is clear. There's actually some solace in understanding this trajectory, even as we navigate the uncertainty of exact timing.
The world will look very different, and it's something I try and prepare for with intention.
Phenomenal! I will be using some of these tools to enhance my life, but also as a guide to set my daughter up for a brighter future
This is awesome, thank you sharing!