AI, Labor, and the Future We Can Shape

Last week, I had the opportunity to join a meeting at the Department of Labor focused on AI and the future of work. It was one of those rare moments where technologists, policymakers, and labor experts were in the same room, wrestling with the same big questions. 

I was very much impressed by the work the Department of Labor is doing to understand the challenges facing AI builders, expand AI literacy, analyze AI’s impact on the labor market, and pilot new models to help workers succeed in the AI-driven economy. 

As I flew back to Seattle and reflected on the conversation, I kept coming back to one thing swirling around in my mind: the people involved in building AI, particularly those who are not technologists but low-paid, hourly workers.  The U.S. quite simply does not currently have enough people to win the AI race, long term. This applies to the highly skilled workforce as well as the frontline workforce. But we will change this, and we will win.

We talk a lot about algorithms and models, but we rarely talk about the people behind them. Every AI system, no matter how advanced, is ultimately trained by humans and the content they produce. Many of those humans are low-paid workers abroad, performing data labeling, content moderation, and reinforcement learning tasks.

Take ChatGPT. Its development relied on offshore labor earning as little as $1.32 per hour to filter toxic content and help train the model. This “training layer” is just as critical as the model architecture itself, yet it’s largely invisible in public discourse.

That’s a missed opportunity. We could create good jobs for American workers in this training layer, roles that would deepen our domestic AI expertise while raising labor standards globally. This has lots of ramifications for immigration policy.

Right now, we’re not moving fast enough. And I believe the DOL gets that.

AI’s impact on labor won’t just be about replacing certain tasks; it will be about reshaping the entire labor market, from what skills are in demand to where critical work is performed. The U.S. needs a clear strategy not only to deploy AI, but to make sure our workforce – all of our workforce – has a meaningful part in building it. 

In my view, the AI revolution is more consequential than the advent of the Internet. 

It will help define global power structures for generations. Sitting back is not an option.

The conversations in D.C. left me extremely encouraged. Leaders at the Department of Labor, including Chief Innovation Officer Taylor Stockton and Chief Economist Julia Pollak, showed genuine openness to hearing from the startup community and engaging with the realities of AI development. 

We have a chance to build a more competitive, inclusive, and worker-centered AI future. It will take far more than a single CEO of a small startup in Seattle to make that happen. But with the right coalition of voices, spanning government, industry, and labor, we can shape an AI era that strengthens both our technology and our workforce.

The question is whether we’ll move quickly enough to do it.