No, AI is not dangerous for humans!

I – US hype

The media are chanting the mantra-like chants of American artificial intelligence, which recently would perform wonders, to the point that it would become a danger to man. The famous Elon Musk, Stephen Hawking, Raymond Kurzweil, all reputed to be (very) intelligent people, make headlines by predicting the domination of the world by a nasty AI. Musk created the company OpenAI to try to prevent it. The famous astrophysicist Hawking proclaims soberly « I think that the development of a complete artificial intelligence could put an end to humanity ». As for Kurzweil, director of engineering at Google, he even gives a date to our collapse: 2045, end of the human era in the face of an artificial « superintelligence ».

Artificial intelligence would already beat the world’s best players in chess, Go, Poker, Jeopardy game show. It would spread in commercial software, in autonomous cars, in voice recognition (Siri), etc.

Do not worry, it’s all American daring complacently relayed by media in search of sensationalism. There is already no AI in Siri (the voice recognition of Google) and Google does not even claim it. Moreover, since forty years that we are talking about the wonders of artificial intelligence, it has still not entered our homes …

II – The symptoms of deception

This hype originates from large US groups like Google, Facebook, IBM, Apple, etc. Hungry for sensational media blindly make their ad, dictated by so-called « IA » researchers (Yann Le Cun for Facebook for example, is actually a specialist in image recognition and machine learning). Thus freed from all control, they can intoxicate us, pretend that this technique is a very difficult goal to reach, that the intelligence is indefinable and that they are the only ones to master the subject. In fact, the « subject » they are trying to control is us …

Indeed, why talk about artificial intelligence without ever defining intelligence? Why despite the countless advertising effects in 40 years does the public still understand nothing about artificial intelligence? Why is it necessary to produce a specific AI by subject when intelligence is deemed to be general? Why is an AI capable of playing chess unable to play Go? Why do so many computer scientists and man-years of development need to program one of these AIs while obviously if a machine is intelligent it does not need a computer programmer to program? Why are so many computers needed to run these American AIs while in France much more sophisticated AIs were shooting in 1982 on the first PCs? Why compare these laborious efforts with human intelligence that fits in a little skull and adapts to anything in seconds? Why are they still computer scientists who speak (very badly) of AI and never its users? Why do we admit since 1950 that the first demonstration of a real AI will be able to dialogue with us in order to make us believe that it is human (the famous « Turing test ») and that no major US university or any of these « familiar » multimillionaire groups of AI did not succeed in 67 years?

III – AI gaming software that cheat to win

You will say: maybe it’s fake AI but it beats the world’s greatest players, you need intelligence behind! In fact to win, these programs cheat. They divert the problem, compensating for their lack of intelligence by something that the human knows very badly and the computer, with its unlimited memory, very well: calculate. For hours, at the speed of the electron, he evaluates an infinity of possible strokes. It’s a lot longer, it’s stupid, but it gives accurate results. But this is no longer play! Literally, because our games use reasoning. We have fun reasoning better than the opponent, that’s what connects us. To calculate does not entertain us and besides, our brain is not made for that. It is made to simulate the universe. It’s already not bad … By calculation it would be an impossible task. By reasoning, we can achieve satisfactory results with parcel data.

IV – AI is not the privilege of computer scientists!

No, AI is not a domain reserved for computer scientists, on the contrary. It is within the reach of the general public. Why ? Because the intelligence is in each of us. Just do a little introspection. Intelligence is reasoning on knowledge. The more knowledge you have, the more intelligent you are. The reasoning was described 2,400 years ago by Aristotle: the « syllogism ». It is a logic that all living things, including plants (read Stefano Mancuso) possess. Without her, we are dead. The man reasons better simply because he has more knowledge, he has more knowledge simply because he has more memory and he has more memory simply because he has a bigger brain. It’s « Logical »!

V – The myth of evil artificial intelligence

The US AI being false and strictly identical to other computer programs it is not more – or no less – dangerous than them. The real one exists in France, but it does not meet any support on the part of the computer world which fears losing its place. Yet associated with the virtues of the computer – unlimited memory, immortal and ultra-fast, constantly increasing knowledge accessible to all, total reliability of results, attentive 24/24, ignoring hesitation and stress – such intelligence will end necessarily to exceed that of an individual. But at his service!

To oppose the man would require an artificial ego. Do not confuse intelligence and ego as do our prophets of doom. Intelligence has no project. She is neutral. It is a mechanic used to solve problems. She does what is said to her without a qualm. It is not from her that danger comes. The ego, he has a project. He seeks the satisfaction of his needs. He wants to cum. For this he needs his territory and fight against competitors who have the same goals. He needs intelligence.

Producing a dangerous AI for humanity raises huge technological problems, would take years and lead its designers into jail for uninteresting results: what would be the use of years of self-sacrifice to develop an AI with a fierce artificial ego that, in principle, will no longer obey its creators and seek to destroy them? What interest ? We already have our neighbors …

Finally, if someone does it anyway, his « super-intelligence » would be facing the most ferocious predator in the history of the planet, equipped to kill, organized to kill, who likes to kill: the man! Billions of men. Poor little beast! His fate is sealed.

Sleep on, you brave folk.