The Four-Minute Mile Was Never the Story

Years ago, I wrote about the four-minute mile.

For decades, it stood as one of sport’s great psychological barriers. People debated whether the human body was even capable of running that fast, though Roger Bannister himself always believed the barrier lived more in the mind than in the legs. Then, in 1954, Bannister did what many people believed could not be done. Once he crossed the finish line in under four minutes, something remarkable happened. The barrier did not simply fall for one extraordinary athlete. It fell for everyone. Within weeks, someone else had done it. Then another. Today, elite runners no longer dream of breaking four minutes. They expect to.

On Saturday, Josh Kerr set a new world record when he ran the mile in 3:42.66.

Think about that for a moment. The question is no longer whether a human can break four minutes. The world’s best runners have moved more than seventeen seconds beyond what many once believed was the limit of human performance. The record Kerr broke had stood for twenty-seven years. Progress like this does not arrive on a schedule. It comes in long plateaus and then, suddenly, in leaps.

The four-minute mile was never really the story. It was only the beginning of the story. The breakthrough gets the headlines, but what comes after changes the world. I think artificial intelligence is approaching that same moment.

Only a short time ago, lawyers and judges were asking whether artificial intelligence could summarize a deposition, draft a brief, prepare a bench memo, or conduct legal research. We debated whether the technology was capable of doing meaningful legal work at all. Could it actually do the work, or was this just another technology industry promise that would never quite live up to the hype?

Today, those questions are beginning to fade. The better systems can already perform many of those tasks. Sure they still make mistakes and still require human judgment. But capability is no longer the most interesting question because, like the four-minute mile, the barrier itself has been crossed. The harder questions come afterward, and they are no longer about capability. They are questions about judgment, governance and about how we want our institutions to function once the breakthrough has become ordinary.

The legal profession has a habit of spending too much time debating whether a technology works and not enough time preparing for what happens after everyone accepts that it does. Every week seems to bring another headline announcing that AI passed another exam, drafted another document, completed another professional task, or reached another benchmark. Those stories still matter because breakthroughs always matter. But they are increasingly telling us something we already know.

The more interesting question is what comes next.

Josh Kerr’s race is remarkable not simply because of the time on the clock. It reminds us how quickly yesterday’s impossibility becomes today’s baseline. Roger Bannister did not change the future of running because he broke the barrier. He changed it because, once the barrier fell, runners stopped measuring themselves against it. They started asking a different question. And we should too.

We have spent the past few years asking whether AI can do the work. That was the right question when the barrier still stood. It is becoming the wrong question now that it has fallen.

The real challenge is no longer proving what AI can do. It is deciding what lawyers, judges, and courts should do once everyone knows that it can.