By akademiotoelektronik, 18/01/2023

Digital immortality, the "delirium" of transhumanists?

Simulate our mind, a dream in the process of achieving?Let us be reassured, Johnny Depp will not contaminate our electrical networks as in transcendence.Whatever Ray Kurzweil and Google says, implanting the neuronal cortex in a machine is still far from being acquired.In this regard, Jean-Claude Heudin, expert in artificial intelligence, had published "digital immortality: artificial intelligence and transcendence", where he strongly criticizes the analogy which is constantly made between the computer and the brain.

While Technoprog, the French Transhumanist Association (AFT), defends the fact that we have succeeded in reproducing networks of neural, Jean-Claude Heudin proclaims that we must stop thinking that digital immortality is possible.We therefore interviewed him to find out a little more about advances in the simulation of the human brain and the progress of artificial intelligence.

Director of the Internet and Multimedia Institute, Jean-Claude Heudin is a teacher and scientific researcher, specializing in artificial intelligence.He co -founded a company specialized in AI and participated in several important applications of artificial intelligence for defense and industry until 1995.He was also a scientific advisor for the City of Sciences, organizer of international conferences "Virtual Worlds" and expert with the European Community for projects "Future Emerging Technologies".

It is always the crystal ball for a subject like AI, which has always been subject to more or less miraculous prophecies.Already in the 80s, it was said that computer science was over and that there would be only artificial intelligence.Artificial intelligence is as old as IT.It is impossible to predict whether in 2024, the AI market will reach billions and billions of dollars.There are things that are part of the AI currently and that may no longer be AI within 20 years.

After that, we are rather in the top of the wave today.There have been recent advances in laboratories with high investments from large American companies such as Google or Facebook.We can assume that there will be long -term benefits.Nevertheless, from there to go in transhumanist delusions, I will not go so far.In the end, the domain of AI is similar to that of robotics, they both arouse fantasies carried by popular culture, which come up against the reality of the laboratories.

Stephen Hawking's thought is interesting but we have to make the share of things.Whether it's the biggest fear today, you have to stop.There are much greater threats than the hypothetical advent of an artificial intelligence that would decimate us in 2045.Currently, we have the possibilities of producing very competent AI in restricted areas, therefore creating general artificial intelligence, it is completely out of our reach.

It's like imagining the Deep program to become an omniscient AI capable of learning on its own.Deep Mind is a human victory before being an AI victory, there is a team of 20 researchers each time for maintenance, it is not autonomous.This remains very specific algorithms so it should not be deduced that man is obsolete and even less the fact that an AI will decimate us all.

The biggest project today is the Human Brain project, a project that has emerged in the United States and is responsible for simulating the entire human brain.Even if we manage to simulate several million neurons, we are still very far away.The vast majority of artificial neural networks are based on layer systems, which do not have the same organization at all as in the human cortex.For example, if you come back to Deep Mind, it has 13 layers of neurons, which allow information to enter per instance.For the moment, the models of neurons made are relatively simple.

If it is the transhumanist project as it is carried today by Ray Kurzweil, Google and its French equivalent, I think we are in delirium.In my book, I precisely take the example of consciousness: we are not ready to make a conscious machine for the moment, so imagine that we can simulate our mind in a machine, it is that weis totally in the imagination of science fiction.

We forget that the brain is connected to the rest of the body and that the body has been shaped by the environment.Suppose our brain is placed in a computer, it's a real nightmare!If we were technically capable of it, we would lead to a being with serious mental illnesses.If you are no longer in contact with your body, you no longer have perceptions.

The brain is often considered the most complex object in the universe and one cannot reduce intelligence to the capacity of calculations.We are shown by the Moore law permanently.All these progress is not exponential, contrary to what Kurzweil says.

We have developed electronic pilot projects related to French Defense.The idea at the time was to have an electronic co-pilot that helps the fighter pilot.A bit like Iron Man and Jarvis.AI is very anthropomorphic in the movies, but it's already almost that.It does not replace the pilot, this is to help him even if today, some are considering completely automatic planes.

Today, many projects are born around digital immortality.Supported by the European Union, Human Brain Project aims to copy the entire human brain through a supermarket by 2024.The goal ?Develop new medical therapies around neurological diseases.The Department of Electrical Engineering in Computer Science and the Institute of Neuroscience Helen Wills of the University of California have proposed to base themselves on nanometric particles (also called "neurons dust") in order to integrate one day into the human cortexWireless interface Brain-Machine, which would keep for life.And a Russian millionaire, Dimitri Itskov, even plans to create an artificial brain to become immortal, by 2045.

It remains to be seen if we can reach the basics of digital immortality in the coming years.In any case, we cannot deny that the human brain is too sophisticated to be reduced to a central unit.And if we stick to Hebb's rule, two neurons are excited jointly and there is a link united them.It is the repetition of the two stimuli that creates associative learning.By repetition, the brain reaches perfection unlike the computer which will not use repetition to set up a task.While it is possible to locate the seat of "thought" in a machine, the thought of the human brain is widely distributed.100 billion neurons are activated at the same time, each of them being connected to 10,000 others.

As Michio Kaku expressed it (professor of theoretical physics at the City University of New York) in "A brief history of the future", the brain is "more analog than digital" and cannot work in binary.In short, digital immortality is not for now.