Why AI is changing everything

Cristina Criddle (0:02)

All engine running, that's the genius.

Chris Smith (0:04)

Get this.

Cristina Criddle (0:05)

Welcome. This is the show where we bring you science.

Chris Smith (0:07)

What that essentially means is discovery is the question. Research.

SPEAKER_3 (0:11)

Technology. Unbelievable.

Chris Smith (0:13)

Without further ado, this is The Naked Scientists. Hello, welcome to The Naked Scientists, the programme that brings you the biggest breakthroughs and talks to the major movers and shakers in the worlds of science, technology and medicine. I'm Chris Smith, and today, we're taking a closer look at the evolution of artificial intelligence. We're calling it YAI and we've been speaking to some of the biggest names in the business.

Today, we're introducing a new series in partnership with the open science publisher Frontiers, and it's called Y. In this series, we'll explore some of science's most pressing questions, from geoengineering to open source publishing and public trust. It comes ahead of the Frontiers Science House in Davos, which brings scientists, policy makers and business leaders together, so open, evidence-based science discussion can help inform global decisions on health, the planet and emerging technologies.

We begin with the big question on everyone's mind right now, which is why AI? In just a few years, this has gone from the realms of science fiction to something many of us use every day. But concerns have grown so much that artificial intelligence is now front-page news and the White House has unveiled a super intelligence force to oversee AI policy. We'll get to that later. But first, how exactly does AI work? One of the pioneers in this space who won the Nobel Prize in 2024 for developing the architecture upon which a lot of this technology is based is Geoffrey Hinton. His inspiration in the early days during the 1970s was the brain and how networks of nerve cells forming stronger and weaker connections among themselves can be modelled in computer code.

Geoffrey Hinton (2:11)

At the time, the idea of doing science by simulating things on computers was fairly new and it seemed just like the right approach to trying to understand how the brain learns.

So, I spent my time writing computer programs that pretended to be networks of brain cells and trying to answer the question, how should the connections between brain cells change so that the collection of brain cells hooked up in a network can learn to do complicated things like, for example, recognise an object in an image or recognise a word in speech or understand a natural language sentence. So, we now have computer models of neural networks that work really well. You see that in these large language models and in the fact that your cell phone can recognise objects now. It can also recognise speech. We're not quite sure exactly how the brain learns, but we have a much better idea of what it is that it learns. It learns to behave like one of these big neural networks.

Chris Smith (3:12)

If it's down to the fact that we've got brain cells talking to brain cells and they're just big populations of connections, is that not relatively easy to model with a computer? What's the hold up? Why is it hard to do this?

Geoffrey Hinton (3:25)

Well, the tricky thing is coming up with the rule about how the strength of a connection should change as a result of the experience the network gets. So for example, very early on in the 1940s or maybe early 1950s, a psychologist called Hebb had the idea that if two neurons, two brain cells fire at the same time, then the connection between them will get stronger. If you try and simulate that on a computer, you discover that all the connections get much too strong and the whole thing blows up. You have to have some way of making them weaker too.

Chris Smith (3:58)

I love that line, what fires together, wires together. It's never left me because I remember reading Hebb's book when I was at University College London. So how did you try and address that then? Was it just a damping problem?

You make it so that the nerve cells get bored more easily as it were, so that doesn't overheat in the way that the computer would otherwise have them do.

Geoffrey Hinton (4:20)

Well that's kind of the first thing you think of and you try that and it still doesn't work very well. So the problem is, can you get it to work well enough so that it can do complicated things, like recognise an object in an image? Or in the old days, recognise something like a handwritten digit. So you take lots of examples of 2s and 3s and so on and you see if you can make it recognise which is a 2 and which is a 3 And it turns out that's quite tricky.