Why Nuclear War Is More Likely Than You Think (with Carlo Rovelli)

Rory Stewart (0:09)

Welcome to The Rest Is Politics for a new series on Existential Risk. And I'm joined today by somebody I admire immensely, Professor Carlo Rovelli. I hope many people listening will already be familiar with his works because he writes very, very beautifully about physics and the universe. Some of the books that you may have come across, Seven Brief Lessons on Physics, Reality Is Not What It Seems. It's a lovely title. I like Reality Is Not What It Seems. But today he's here actually engaging much more in something much more political and he's here to talk to us about nuclear war. I think I'm going to begin you there, Professor, and then we'll come back to talk about who you are and bring out some of this wonderful lively mind. But let's begin with that. Nuclear war feels a quite unfashionable subject. It feels like something that my father-in-law and mother-in-law were worried about in the 1960s. They were all building bunkers. They all thought the Cuban Missile Crisis was coming. Everybody was writing books. The end of the world is nine, nuclear war is happening, and then it didn't happen for more than 50 years, 60 years.

So people maybe aren't talking about it quite as much as they were. Is this because the risk has changed?

Why are we talking about it less than we were, do you think?

Carlo Rovelli (1:33)

Thanks for having me, Rory, and thanks for this very kind introduction. I think that what you just said is the source of the risk and the reason for me for writing this book about the risk. Namely, it does sound unfashionable, doesn't it? Just think of the 60s, when there were atomic refuges and things like that. We're over that, and that's exactly the problem, because objectively, and I use this word with a lot of caution, but I think it has to be used. The risk is higher than the 60s.

The title of my book is 85 Seconds to Midnight, refers to the slogan used by the Bulletin of Atomic Scientists, which is a committee of scientists that tries to use objective parameters to evaluate the risk of a nuclear catastrophe. And they come out with the conclusion, they do this every year since the Cold War, and they came out with the conclusion that now the risk is higher.

Rory Stewart (2:40)

What are the factors which might make it more likely today than it was in the 60s?

Carlo Rovelli (2:44)

There are several, at least four. Let me make a short list of them. First of all, what you said at the beginning, talk about atomic bombs looks out of fashion. Nobody's talking about that.

People are forgetting the essential risk that they imply. The political discussion in many countries, especially in Europe, about rearmament, it doesn't include the part, yes, and then if there is war, this war is nuclear.

The nuclear issue is not in the center of the discussion as it should and as it was in the 15th and the 60s. So that's one reason. The second reason is that it's not anymore two superpowers. The world has changed, the Cold War has ended, Soviet Union gave up, essentially. The Soviet leadership decided we can become capitalists like you. It's better for us. Now there is a third nuclear superpower which is coming out, which is China. And it's not clear what Europe is going to do with its French and British nuclear armaments, which are far less than the other three. Game of three is completely different to game of two, because the game of two is you watch one another and we agree not to go disaster. Game of three is completely different because there's a complicated thing about trust and there is a clear advantage in everybody having the other two going to war. So somebody is pushing always for the other two going to war.

That's the second reason. The third reason is artificial intelligence, because not using artificial intelligence in a game, which may be a game of minutes, very fast decisions. Remember, now we have the equivalent of nuclear bombs in Turkey. Nuclear bombs are very close to Russia right now. So AI, it's likely to be used in combination with nuclear technology, but AI is not fully reliable in these things, because the high point of AI takes decisions faster than humans. So you don't want to slow down it with humans. And then you put your hands in the decision of AI system, and we know that they are completely unreliable on substantial things today.

Rory Stewart (5:08)

And the interesting question is, with AI is so manyfold, and we can get into them in a second, but there's a question of unreliability, hallucination. There's also a question of rigidity. There's a question of these models being trained to follow a particular form of logic, which may be exactly the logic that the military advisors of Kennedy and Khrushchev would have pushed for, which leads to war as opposed to the strange forms of political imagination, which allowed the three people to say that.