Is AI... Physics? This Year's Nobel Prize Raises Questions, But Winner Hopfield Answered Years Ago

This year’s Nobel Prize in Physics, awarded to John J Hopfield of Princeton University and Geoffrey E Hinton of the University of Toronto, has sparked lively debates across scientific circles and social media alike.

While the recognition of their work in artificial neural networks and machine learning has been groundbreaking, some critics argue that the prize didn’t exactly go to an accomplishment within ‘traditional’ physics.

But does physics have to stay within those confines? And how do these discoveries align with the discipline’s broader goals?

Physics Meets AI

Hopfield and Hinton were awarded the Nobel "for foundational discoveries and inventions that enable machine learning with artificial neural networks."

These networks, a bedrock of modern artificial intelligence (AI), are not generally considered subfields of physics. Yet the laureates employed fundamental principles of physics in their research — particularly tools from statistical physics and condensed matter physics — to make strides in machine learning.

The Nobel committee probably knew that the physics connection would come into question, or else the popular science background information on the website wouldn't carry a headline that reads like a clarification: 'They used physics to find patterns in information.'

As the Nobel Prize website notes, Hopfield’s famous "Hopfield network" draws from physics concepts used to explain a material’s characteristics through atomic spin, a property that makes each atom act like a tiny magnet.

Similarly, Hinton applied tools from statistical physics to develop the Boltzmann machine, a model of neural networks designed to learn from its environment.

So, while AI and neural networks are typically categorised under computer science, these laureates worked at the intersection of disciplines, borrowing from physics to come up with breakthrough technology.

A Blurred Line

Naturally, awarding the physics prize to AI researchers has stirred some dissatisfaction. Shouldn’t the prize have gone to advances in physics as we all know it to be, like particle physics or cosmology? The blurred line between disciplines, however, is not exactly new in science.

"In physics, we use artificial neural networks in a vast range of areas, such as developing new materials with specific properties," said Ellen Moons, Chair of the Nobel Committee for Physics, as quoted on the Nobel website.

So, while the laureates' work may not be traditionally categorised as physics, it continues to have far-reaching applications within the field.

As for the debate? It will likely linger. But one of this year’s winners, John Hopfield, has a compelling take on the question: What exactly is physics? He presented this take several years before winning the prize.

Physics As A Way Of Thinking

For him, physics is not limited to subject matter like atoms or galaxies. Instead, it’s a “point of view” rooted in understanding the world through observation, experimentation, and quantifiable conclusions.

“Being a physicist is a dedication to the quest for this kind of understanding,” he writes. It’s an attitude toward solving problems that transcends specific fields, whether those problems concern the nature of matter, biological systems, or — as is now the case — AI.

Hopfield’s career trajectory is a testament to this interdisciplinary mindset. Early on, his curiosity about the physical world — whether it was fixing bicycles or flying model planes — led him to study solid-state physics, which was considered more akin to chemistry than other branches of physics at the time. (Read the description of this textbook on solid-state physics.)

From Solid State To Biological Physics

Solid-state physics may have been a newcomer in the mid-20th century, but it quickly gained prominence due to its role in developing technologies like transistors and integrated circuits — the cornerstones of modern electronics.

As solid-state physics expanded into condensed matter physics, its principles also began to apply to biological systems. “Biological physics began when well-known physicists who had become interested in biology… began pursuing this interest while keeping in strong touch with their physics roots,” Hopfield notes.

In fact, his foray into biological information processing — a field seemingly far removed from solid-state physics — was, in his words, “entirely accidental.” His move into studying the nervous system and neurobiology opened the door to the contributions that eventually earned him a Nobel Prize.

Despite his drift from conventional physics departments, Hopfield always felt that his research remained “entirely in the spirit and paradigms of physics.”

Bridging The Disciplines

By the time Hopfield joined Princeton University’s Molecular Biology Department (a return to the university) in 1997, he had firmly established himself as a scientist who defied categorisation.

“Although no one in that department thought of me as anything but a physicist,” he recalls, “there was a grudging realisation that biology could use an infusion of physics attitudes and viewpoints.”

His wide-ranging career, which touched on everything from solid-state physics to neurobiology, exemplifies the notion that physics is not bound by traditional definitions. To Hopfield, it’s all part of the same quest to understand the world — and to him, that’s what physics is ultimately about.

Hopfield’s career serves as a reminder that the boundaries between disciplines are often porous. In fact, some of the most significant discoveries happen at these intersections.

As Hopfield says, “Physics is a point of view about the world.”

Fair enough! But a Nobel Prize in physics for solving pure, often hard-as-hell physics problems would be nice, wouldn’t it? Conventional physicists might argue in return.