L’Écho – Paul Jorion: “With AI, we have created something that is more intelligent than we are”, March 5th 2024 – Online version


Paul Jorion: “With AI, we have created something that is more intelligent than we are”.

Paul Jorion has been working on artificial intelligence at British Telecom since the 1980s ©BELGAIMAGE

Paul Jorion talks about the moment of technological “singularity”, when AI surpasses human intelligence and even acquires consciousness.

Anthropologist and economist Paul Jorion has already published a number of books. His latest work, published on 6 March, is devoted to artificial intelligence (AI): “L’avènement de la singularité. L’humain ébranlé par l’intelligence artificielle” (1). But Paul Jorion is not just an author. He also co-founded an AI start-up, Pribor.

What is the origin of this start-up, Pribor?

You have to go back a bit in time. In 1987, I was called by the director of the CNRS in France, who said to me: “This is a big scandal, Mr Jorion. You’ve been blacklisted for years, I’m going to put that right”. I was then appointed to the Laboratoire d’Informatique pour les Sciences de l’Homme in Paris.

I went to Bordeaux to attend a major conference on artificial intelligence, a field that was still in its infancy. In the corridors, I was approached by an Englishman who offered me a job at British Telecom, where an artificial intelligence team was being set up. It was there that I developed the Anella software, an associative network with logical and learning capabilities. In 1990, with the end of the Cold War, we were sadly told that there was no longer any funding, because it had come from the army. So I could no longer continue in this field. That was that, until 2021, when some young Italians, who had a small laboratory in Milan, came to see me and told me that what I had done at the time was revolutionary and that it would be interesting to submit a project to the European Union for funding. But apparently the person in charge at the European level didn’t understand what was going on and the money was distributed to projects that were completely out of date. We were totally dismayed. It was at this point that I was approached by a French businessman, Manuel Guérin, active in the oil industry, who had read my blog and said he was interested in the project. So in June 2022 we set up a small company called Pribor.

What exactly is the aim of your company, Pribor?

We were able to recreate entirely what I had done for British Telecom, but in a contemporary context, with more powerful tools, in order to develop Self-Aware Machines (SAM), AI based on a genuine understanding of the human psyche and not on the probability that one word will follow another. It’s making good progress. The company is based in Paris. There are three of us, including a researcher.

Paul Jorion

But in the meantime, I have to admit that progress in AI is quite extraordinary. There was ChatGPT born on 30 November 2022 and then the ChatGPT4 version on 14 March 2023, a sophisticated tool that even had to be restricted. And just recently, there was Sora, unveiled on 15 February by OpenAI: all you have to do is write a sentence and the machine will create a one-minute video. It’s pretty amazing. Someone told me that a friend of his said that when he saw such progress, it was as if his job as an advertising executive had suddenly disappeared.

You write that on 14 March 2023, the release date of ChatGPT4, we reached the technological singularity. Is it the machine that will definitively overtake man?

Yes, this is the moment in history when artificial intelligence makes exponential progress, rapidly surpassing human intelligence, thanks to its ability to programme itself.

Because the machine is more intelligent, it doesn’t need humans to perfect itself. We’ve already seen this in games like Go. All the specialists, even the computer scientists, pointed out that the number of possible combinations in this game exceeded the calculation capacity of a computer. But AlphaGo beat the various Go champions by studying the games played by humans. Its successor, AlphaZero, learned Go from the rules of the game alone. And it beat AlphaGo by 100 to 0. By putting humans in brackets in this way, the machine was even stronger.

You write in your book that the machine has a conscience. Luc de Brabandere, mathematician and business philosopher, maintains that a computer can simulate discomfort, but it cannot feel it. It can imitate emotions better and better, but it remains above all a pile of metal or plastic parts. A machine will never be able to come up with a good title for a novel,” he says.

That was all true until the arrival of ChatGPT 4. Since then, things have changed radically. Someone on my blog showed me a poem by Prévert, arguing that the machine would never be able to match it. So I explained to the machine how Prévert did it. And I can assure you that on the third try, the poem was better than Prévert’s.

Whether the machine already has consciousness – as I am personally convinced it does – or whether it acquires it in the months or years to come, it is already giving us access to a wealth of knowledge that will multiply our ability to think about any problem. Geoffrey Hinton, one of the pioneers of AI, was astonished himself. He suspected that we were going to create machines that would do as well as humans. But he never thought it would happen in his lifetime. Today, he recognises that the machine is indeed superior to man.

In terms of the opportunities offered by AI (medicine, the environment, etc.) and the dangers it poses, what is the right balance?

I have the impression that with the problems of the environment, resource depletion and global warming, we need major technical solutions. And AI offers them.

Artificial intelligence will also enable us to make decisive progress in healthcare. Since 2022, AI has made it possible to predict the way in which complex protein molecules fold back on themselves, enabling us to know what the possible interactions will be between them and other molecules, a type of knowledge that is imperative in pharmacology. The task was considered titanic, but it was solved by AlphaFold, a product of DeepMind, a subsidiary of Google.

The jobs most likely to be downgraded by AI are white-collar ones. What’s the solution?

Yes, it’s intellectual workers and creators who are most at risk.

I reiterate my proposal for a tax on robots, the Sismondi tax, named after the Swiss philosopher and economist. The wealth created by the machine should be subject to a Sismondi tax, which would make it possible to ensure that the things that are essential to the entire population, whether housing, food or local transport, are provided free of charge. The time has come to introduce such a tax. It would be a cheaper solution than a universal income. A British study has shown that providing the essentials free of charge, which could be called a basic universal service, would cost three times less than a universal income.

In philosophical terms, can we say that the machine has become in some ways the equivalent of God?

We have created something that, in our stories, was not science. A being whose intelligence surpassed ours could only be described as God. But until now, we thought that we could never do better than God. Today, we have achieved something that only gods could do: we have created something that is more intelligent than we are. And it is humans who have achieved this feat, I would like to emphasise.

What is significant is that a pioneer like Geoffrey Hinton seems to be afraid of what he has created. He resigned from Google to reflect on his creation, rather like Oppenheimer did with the atomic bomb. There’s also the intellectual Douglas Hofstadter, who says he’s totally discouraged and gives up when he realises that the machine is superior to him. I find these remarks a little sad and desolate. For my part, I see this as extraordinarily exciting because of the opportunities that are being created.

L’avènement de la singularité. L’humain ébranlé par l’intelligence artificielle, by Paul Jorion. Éditions Textuel. 128 pages, 14,90 euros.

2 responses to “L’Écho – Paul Jorion: “With AI, we have created something that is more intelligent than we are”, March 5th 2024 – Online version”

  1. It’s becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman’s Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher order consciousness, which came to only humans with the acquisition of language. A machine with only primary consciousness will probably have to come first.

    What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990’s and 2000’s. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I’ve encountered is anywhere near as convincing.

    I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there’s lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.

    My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar’s lab at UC Irvine, possibly. Dr. Edelman’s roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461

  2. You say that

    A machine with only primary consciousness will probably have to come first.

    That’s the way evolution proceeded: primary consciousness, then later higher order consciousness. But as you rightly say, higher order consciousness “came to only humans with the acquisition of language”.

    There is no doubt that Edelman assumed that primary consciousness was necessary to lay the foundations for higher order consciousness, generalising from evolution. However, what the emergence of higher order consciousness in LLMs fully displays is that primary and higher order consciousness can emerge independently as LLMs, through the acquisition of language at the level of sophistication attained by contemporary humans, went straight to higher order consciousness.

    Claiming that higher order consciousness can only be achieved if primary consciousness had arisen first is unwarranted generalisation of what has been observed in evolution.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.