Who wins artificial intelligence or humans? A question that may be very strange, and is looking for an answer urgently.

On May 11, 1997, the machine proved that it could outperform humans in complex tasks, but it was unable to keep up with them, in terms of the ability to adapt and carry out various tasks, after the Deep Blue computer won over chess champion Garry Kasparov, according to a report by the Emirati newspaper Al Bayan.

Artificial intelligence specialist, Philippe Roulet, told AFP that the victory of Deep Blue was a “culturally surprising” event, but “it was just a good success from a technological point of view.” The co-founder of “Artifact”, a company specialized in consulting and technological development in this field, with 800 employees, explains that the success of “Deep Blue” did not implicitly represent anything but a victory for the “savage force”.

The machine achieves successes in chess, thanks to its enormous computing power, which allows it to determine all the possibilities associated with the development of play during a match, and to deduce the moves that allow for victory. While Deep Blue has mastered the game of chess, thanks to a complete structure of logical rules, developed by humans, the new machines themselves set their own rules, within a learning period during which they absorb huge amounts of data. Philippe Roulet says that what is happening constitutes a “transition from imposed programming to programming by learning”.

The impact of programming on learning in games is very important, as the AlphaGo algorithm, affiliated with DeepMind, which is based on machine learning, defeated in 2017, the best player in the world, the Chinese KJ.

The French MP and mathematician, Cedric Villani, who prepared a foundational report on artificial intelligence in 2018, explained that the defeat inflicted on humans by the creation of “AlphaGo” proved that “the human being is less ingenious than he thought. It turns out that some details of the algorithm, which were initially thought to have been elementary errors, are in fact very important achievements.”

A general method of self-learning The French startup, Nokai, recently explained that machines are now even capable of defeating humans, in games that exist in mysterious worlds, such as poker and bridge. These very important developments go beyond the gaming world, as the device no longer has many achievements to achieve. The head of research related to artificial intelligence at the company “Meta”, and one of the most prominent founders of modern artificial intelligence, Jan Lukun, told AFP that artificial intelligence has achieved “amazing progress, which came as a great surprise” to him over the past years.

“We are currently able” to allow a machine to “translate any language into another, within a group of two hundred languages”, or “to have a single neural network that contains about a hundred languages,” he adds. But machines still run into obstacles The possibility of having an interesting conversation with GPT3, a script generator developed by startup OpenAI for its billionaire owner Elon Musk, does not mean that the program “will provide us with assistance in everyday life.” According to Yan Locon, the missing step in developing this virtual assistant, and in making a truly autonomous car, is to reach a “comprehensive self-learning method”.

And he adds, we put the computer in front of two hundred hours of video clips, and through this step, it achieves a kind of “understanding of the world” and “good perception”, and this then allows it to reach learning capabilities “similar to those we observe in humans.” animals and humans.”

The researcher expresses his conviction that machines will one day reach “a wide ability to learn, and they will be able to learn everything that humans learn, with superior capabilities in most cases,” adding, “But when will we reach this stage? The answer is not clear.” But the actual revolution is what is being developed in laboratories, which includes machine learning (or automatic), and artificial neural networks, which have made great progress for artificial intelligence, during the past decade.