Has the Nobel Committee jumped on the AI hype train? Reflections on the implications of this year 2024 choices

First, let me emphasise that these reflections on this year’s Nobel Prize winners in Chemistry and Physics are not intended to diminish the remarkable achievements of the awardees. The opinions here expressed are mine only.

Despite the Nobel Academy’s efforts to frame this year’s prizes as achievements in chemistry or physics, this year Nobel Prizes went to works of computer science and AI, even though they contribute significantly to physics and chemistry.

This year choices raise some interesting questions to reflect on:

Are Nobel Prizes now being awarded for theoretical tools that enable discoveries rather than to the discoveries themselves?

As some of you know, I’ve been a long-time critic of the Nobel Prize system (and similar) and for full disclosure, I was associated to one of the awarding institutions for several years as Assistant Professor and Lab Co-leader at the Karolinska Institute. I have also been heavily involved in thinking and writing about AI for scientific discovery at The Alan Turing Institute and in my last position at The University of Cambridge.

While I think recognition is important as a human social need to both the receiving end and those awarding (and praising about) it, I believe that prizes can serve the wrong purposes, bring the wrong incentives (e.g. deforming scientific goals) and we could rather design them as powerful tools to tackle biases and inequalities.

These prizes celebrate highly established figures — many of whom are already well-supported. By the so-called Matthew’s Effect, we tend to give prizes to those that already had prizes in a snow-ball effect. An alternative would be to focus on promoting young researchers of equal excellent work who do need the visibility and the financial backing (which can be split among a larger group). Instead, prizes often create ‘demi-gods’, placing excessive pressure on laureates harming their future work with people paying attention to every word the say, sane or insane, and widening inequality gaps in science and society.

This year’s physics laureate, British-Canadian researcher Geoffrey Hinton was celebrated for his work on neural networks, in particular a method called back-propagation. However, he was not the first to discover the method (e.g. Kelley, Dreyfus, and others) yet, Hinton has been instrumental in connecting past efforts to current solutions, is the most well-known and definitely the most vocal and media covered. Does the Nobel Prize take popularity into account perhaps unintentionally and what are the measures taken to correct this bias and how?

If prizes are necessary, would not make sense to use them to close equality gaps in science?

The underrepresentation of women in this year’s choices is glaring — no women have been awarded any, clearly indicating a challenge of root causes. Meanwhile, many young researchers — often stuck with the demeaning label of ‘postdocs’ — do the heavy lifting in research, surviving on stipends.

If Nobel Prizes are now awarded for tools that facilitate progress in other domains, where are the prizes for pioneers like McCulloch and Pitts, who invented artificial neural networks, von Neumann or Shannon and Turing, who laid the foundations of information and computing among hundreds of others? This year’s awards risk perpetuating historical unfairness rather than ever correcting it.

Should we expect CEOs to start receiving Nobel Prizes if their companies achieve scientific breakthroughs?

Take Demis Hassabis of Google DeepMind, for instance. While I have immense respect for him as visionary and computer scientist (and he is also a very nice down-to-Earth person from the couple of brief interactions I have had with him), as a CEO, he most definitely delegated much of the detailed work, and his attention was likely divided in a wider number of projects. He is possibly the first Nobel laureate in a domain X where his core expertise is definitely not X.

Would Nobel Prizes go to CEOs of companies with the resources to commercialise groundbreaking solutions or to the domain experts on which not only the training data came from but also those who the architecture was designed by? I predicted that Alphafold would get the Nobel Prize about two years ago for bad or good, sometimes voicing it on LinkedIn.

Alphafold is not an agnostic naive system, it was designed by and with input from domain experts for years. Alphafold has probably also not added much, if a single line, of new understanding of the first principles (the science) of protein unfolding to scientific textbooks (if not interpreted by academics using Alphafold) making it rather a technological solution than a scientific breakthrough in the protein folding field.

While these Nobel Prizes may look like modernising an archaic reward system, we need to keep asking tough questions: Are these awards primarily about science, or do they have a strong marketing component that make them yet another popularity contest perpetuating inequalities in science and society?

The Nobel Committee has definitely helped now increase the valuations of the companies associated to the awardees helping them in their well concerted efforts to keep valuations high to keep raising ever increasing sums of money that helped them implement those technological solutions in the first place. If this is clearly not an intended outcome, how does it help widen inequality gaps between academia and industry, scientists on grants versus private financial enablers?

One clear thing did good the Nobel academies since last year in their choice of awardees coming from the fields of complexity science, and now computer science, information theory and other highly theoretical fields, this is that they are totally relevant to the empirical sciences and not disconnected from as some authors have suggested with no basis (read this other post).

I feel some degree of pride in belonging to fields that are now considered worth getting a Nobel Prize when before it was strangely not and there was no way for the Nobel committees to continue ignoring or overlooking them, but I would be more proud if we could discuss openly how to make science fairer in a highly unfair society and focus on more pressing challenges. This year welcomed changes means radical change in ancient institutions is possible.