What will it take for America to lead the global AI competition? Now we all know we’ll need plentiful energy, the best chips, the most innovative AI companies, the best models. But I argue that’s not nearly enough.
Because if we’re the best in the world at those things, but we’ve displaced millions of American workers, then we’re going to lose the global AI race. In fact, we will have automated our decline. Because recessions, social unrest, political upheaval — those will weaken our country, our politics, and our economy. And ultimately, I believe, I know, they will lead to excessive, burdensome regulation of AI that will slow and stop and hinder AI innovation.
The reality is the United States can’t lead the world in technological innovation if it’s failing its people at home. We need a technology strategy and a human capital strategy, because the only way to truly win the long-term AI race is to lead in the technology and to have a plan that brings everybody along to an AI economy.
The Promise — and the Near-Term Disruption
Now I think it’s incredibly exciting to think about what a well-trained individual can do with AI. Think about how much more productive and creative we’ll all be, and how many new businesses will be created. And I’m optimistic, because history shows that every time we create a new technology, it does create new jobs, new industries, new products, new services — over time, with some time.
And I’m confident that that will happen again, this time with AI. With time.
So what I’m worried about is the near-term disruption to workers as we transition from here to an AI economy. Because — I know this as a former governor and secretary of commerce — America’s workforce and career transition systems weren’t built for this moment. Some people estimate that tens of millions of American workers are in AI-vulnerable jobs. All kinds of jobs, people of every age, geography, income, and level of education. We are not prepared for this transition. And Americans know it.
I was at a bar the other night watching NCAA basketball. Huge University of Michigan fan — Go Blue! And it’s all everybody was talking about. It’s all the chatter. “What are you going to do when you lose your job to AI?” “I’m so bummed. I’m paying all this money for tuition for my kids in college. Are they going to have a job when they’re done?”
Americans are anxious for a reason, and we owe them more than empathy. We owe them a plan. We owe them action.
Two Bad Ideas — and What We Actually Need
And so far, I’m not hearing a lot of good solutions. There’s the slow AI, stop AI, overregulate AI crowd. It’s a bad idea. It denies Americans the promise of AI, and China will pull ahead. And then there’s the universal basic income crowd. That’s also a bad idea. Every one of you knows a job is more than a paycheck. It’s dignity, it’s purpose, it’s pride. Without purposeful work, a society unravels.
So what are the elements of an effective transition plan? Unfortunately, as I stand here today, I don’t know the exact details. But here’s what I do know. I know it’s rooted in a new grand bargain between government and business. I know it means tearing down the wall we’ve had for so long between school and jobs. And I know industry — every company — needs to help lead the transition. And here is what I know more than anything else: I know that if we are determined, we can make it happen.
Fixing the Workforce Training System
At a minimum, we need massive changes in both our workforce training system and our career transition support system. In an effective workforce training system, employers define where work is today, what skills are needed, and where it’s going. And then schools and government training programs prepare people to get there. That is not what we have today.
Today, in our country, we spend hundreds of billions of dollars to incentivize enrollment in college — without regard to whether people get a degree, the skills they need for a job, or a job. The truth is, and you all know this, government and schools don’t know the skills employers need today or will need tomorrow. Industry has the most accurate and dynamic view.
By the way, I’ve seen this be successful as governor and secretary of commerce. When TSMC decided to expand manufacturing of chips in America, they told us what they needed. They needed skilled electrical engineers and equipment operators. So we got to work. We designed, with them and with community colleges and certificate initiatives, accelerated certificate programs and apprenticeships tailored to the company’s needs. Today, TSMC is thriving in Arizona, making leading-edge AI chips at scale for the first time in America’s history.
[Applause]
Another problem with our current system is it’s a one-and-done system. You graduate high school, you go to college, you’re done with your education. That isn’t going to work in an AI economy. All of us — all of us — will have to learn new skills, because our jobs will be constantly changing over the course of a career.
Let’s be honest: most people learn most of the skills they use on the job every day, on the job. So why don’t we have more effective, affordable, flexible options at scale, so people can earn while they learn continuously over the course of a career?
A Story From Real Life
A good friend of mine’s husband recently lost his job in IT. Super smart guy in his late 30s. He spent a long time figuring out a job that excited him, one he thought was an AI-resilient career path. He finally found a job in the HVAC industry. He was psyched for it. He spent a lot of time looking for a training program in his area. He found one that he could get into. It took over a year, cost money, and he wouldn’t be paid for that year. Who can go a year without a salary?
Now look — I know, I’ve seen it — there is a lot of really excellent innovation out there around employer-led worker training, apprenticeships, co-ops, and college partnerships. But here’s the reality: all of those efforts are a tiny portion of America’s post-high school system. They ought to be the system. They ought to be the norm and not the exception. And there should be no stigma for going that path.
Reforming Career Transition Support
The reality is it’s going to take more than just a new workforce training system if we’re going to get through this economic transition, because we also have to provide support for people as AI changes their jobs. And unfortunately, today in the US, our primary career transition support system is unemployment insurance. It was created 100 years ago, in a different time when people had a single job in a single industry for decades. It does nothing to support people getting new training, starting a new business, or entering a new field. And actually — I know this from being governor — it doesn’t provide nearly enough income support for middle-class wage earners, let alone high-income earners.
So in addition to unemployment benefits, why don’t we offer temporary wage support to get workers back into the workforce quickly, by topping up their salary if they take a pay cut to enter a new field? By the way, one of the reasons I’m excited about AI is I think it’ll make starting a business easier than ever. So why don’t we look at a program of self-employment assistance to support workers while they start a new business?
Imagining a Better System
So let’s say we do all this. Let’s say we are determined, creative, and we do all this. What might it look like?
Imagine a 45-year-old woman. She’s been an accountant for 15 years, closing the books. She’s got two kids, a pile of bills, and a mortgage — pretty typical. Last week she was told she’s losing her job because it’s being automated. Today, in our system, if she’s lucky, she’d get a retraining brochure and two weeks’ severance. There are over a million accountants in America.
In a better system, her company would be incentivized to allow her to start retraining months before she’s laid off. Maybe she gets a short-term credential to learn a higher-value skill. It puts her in a position to get redeployed at that company, if the company is committed to redeployment. And if she needs it, maybe she could collect temporary wage insurance to top up her income — to make up the difference between her old job and her new job salary.
Incentives, Innovation, and Urgency
So how do we get there? Incentives, innovation, and urgency.
First, the government needs to fund schools and training programs on their outcomes.
[Laughter]
[Applause]
How about that? They shouldn’t just get the money because people show up and enroll. They need to be funded based upon whether people get good skills and actually get a job.
Incentives need to change for businesses as well. Right now, the incentives are such that a company lays a lot of people off today, and their stock price surges tomorrow. It is too easy to hit the easy button of layoffs. Companies need different incentives. Quite frankly, we need a new system where it’s more expensive to abandon workers than to retrain them.
[Applause]
We all respond to incentives. How about we pilot tax credits or other economic incentives that reward companies for worker redeployment, for entry-level hiring, for reinvesting AI productivity gains into new jobs? We have spent decades — if not longer — perfecting the incentives for investing in machines. We need to do the same so companies invest in people.
And here’s the reality: it’s in all of our interests to have a smooth transition to an AI economy. It isn’t business versus workers. Nobody benefits from recession, excessive AI regulations, social unrest, political violence, and divisiveness. By the way, it’s in everyone’s benefit to reach this exciting potential of AI innovation. It’s not corporate charity to do this. Last time I checked, agents didn’t walk into the store and buy things. Humans do that. And they need money in their pocket to do it.
We’ve Seen This Movie Before
So it turns out America has seen a similar movie to this before — when we didn’t plan for an economic transition — and by the way, it didn’t end well. It happened when American companies moved their manufacturing overseas, mostly to Asia, chasing cheaper labor, and millions of Americans lost their jobs. A lot of communities were crushed.
How do I know this? Because I lived through it. In the early ’80s, when the Bulova watch factory closed, my dad lost the only career he’d ever known, after devoting almost 30 years of his life to the job. One job, one company, one identity. Gone. Here’s the thing: he was 56. He still needed a job. He still needed to work.
My dad was a smart guy. He understood the economics of outsourcing to China, and he knew it would help some American businesses grow faster and add new jobs. But he and all of his buddies still needed a bridge to another chapter of work. And there wasn’t one. After a lot of pain and bitterness on his part, my family got through it. But the truth of it is, my country paid a huge price for that poorly planned transition. In fact, I argue, 30 years later, we’re still paying a price for that poorly planned transition — in the form of increasingly divisive, dysfunctional, violent politics.
[Applause]
Think about this: at that time, it was a few million Americans who lost their jobs — some say two million, some say three million. Let’s resolve right here, right now to do it differently this time.
America Reinvents When the Stakes Are High Enough
Here’s the good news. History also shows that when the stakes are high enough, America reinvents. After World War II, record public investment and research seeded new industries. COVID accelerated growth in clean energy and in health care. AI is a 100-year technology and needs a 100-year response, so that all Americans can reap the benefits of an AI economy.
Why am I optimistic? Because if we’re a country that can design the best chips in the world, create the best models in the world, and spend trillions of dollars to build out our AI infrastructure, then we’re up to this challenge.
Because here’s the reality — and you know this. Our future isn’t predetermined. It’s ours to create. It’s why you’re all here. So let’s get to work.
Thank you.
[Applause]