Lee Sedol, the Go champion famously defeated by Google's AlphaGo in 2016, warns in The New York Times this month that AI may rob us of our awe for human creativity and innovation. While I am sympathetic to his position, I don’t agree.
Yes, AI beat him at Go. So what? Trains have outpaced humans since the steam age, yet we still gather in Paris this month to watch humans run. The joy of human athleticism is unmarred by the knowledge that there exist machines that can run faster, jump higher, and lift heavier weights than any human. AI's mastery of Go no more diminishes human achievement than a cheetah's speed makes our 100m dash meaningless.
Lee's anxiety stems from a deeper source: the shocking realization that intelligence itself—long considered the crown jewel of human exceptionalism—is a domain where machines can threaten our supremacy. It's a gut punch to our species-wide ego.
But Lee's claim that AI will sap the joy from human creativity is predicated on the flawed belief that human worth is defined primarily by raw intellectual prowess. It's an understandable position for a Go grandmaster, but it's dead wrong.
Deep neural networks, notably the latest generations of transformer-based foundation models, are becoming scarily good at intuitive intelligence1. They'll only get better. But this forces us to confront a fundamental question: What exactly is the relationship between intelligence and the essence of being human?
For millennia, we've conflated intelligence with humanity. Our ability to reason, to solve complex problems, to create and innovate—these have been our calling cards, the attributes we've used to distinguish ourselves from other species. But as AI systems begin to match and even surpass us in these domains, we're forced to reassess.
The task ahead is twofold and daunting. First, we must undertake the deliberate effort of disentangling intelligence from the myriad other qualities that define us as human beings. This isn't just an academic exercise—it's a philosophical and practical imperative that will shape our self-perception and our societal structures in the coming decades.
Second, we must grapple with the implications of this separation. If intelligence can be replicated (and potentially improved upon) by machines, what does that mean for human value, for the labor market, for education, for our very sense of purpose?
This struggle—to define ourselves beyond our cognitive capabilities—will be one of the defining challenges of the 21st century. It will force us to reevaluate our economic systems, which often prize intelligence above all else. It will compel us to reconsider our educational paradigms, shifting focus perhaps from rote knowledge acquisition to uniquely human skills.
Moreover, this redefinition of human essence will likely be an ongoing process, evolving as AI capabilities expand. We may find ourselves continuously pushing the boundaries of what we consider uniquely human, only to have AI encroach upon that territory as well.
The irony is that this very struggle—the effort to understand and define ourselves in the face of AI advancement—is itself a profoundly human endeavour. Il faut imaginer l'humanité heureuse. It requires not just intelligence, but introspection, philosophical inquiry, and a deep reckoning with our place in the universe.
Lee Saedol's defeat at the hands of AlphaGo wasn't the end of human creativity or value—it was the opening salvo in a new era of human self-discovery. The game ahead isn't about outcompeting AI intellectually. It's about understanding what we are beyond our intelligence, and using that understanding to chart a course for humanity in an AI-augmented world.
This is the real challenge that Lee's loss illuminates. And it's a challenge that, by its very nature, is profoundly human.
Our best models still struggle with reasoning, but progress is coming fast and there is no reason to believe we will not be successful at deeply integrating formal reasoning and the intuitive cognitive computation of today’s foundation models.