Artificial intelligence (AI) threatens humanity with possible “extinction”, warn several experts and entrepreneurs in this booming sector of activity, and call for awareness of the risks. But is this catastrophic and far-fetched scenario believable, asks the AFP agency, quoted by Agerpres.

This nightmarish situation, inspired by many sci-fi movies, will begin when machines see that their capabilities are superior to humans and get out of control.

“From the moment we have machines trying to survive, we’re going to have problems,” Canadian researcher Joshua Bengio, one of the “fathers” of machine learning, said recently. He also urged governments to act quickly and develop regulations to govern AI development

According to the version proposed by the Swedish philosopher Nick Bostrom, the decisive moment will come when machines know how to produce other machines by themselves, which will cause an “explosion of intelligence”.

According to the “paperclip theory”, if artificial intelligence set, for example, as the ultimate goal to optimize the production of this office accessory, it would eventually cover “first the Earth, then larger and larger areas of the universe, from a paperclip,” he explained.

Nick Bostrom is a controversial figure after claiming that humanity may be a computer simulation and after supporting theories close to the eugenics thesis. He also recently had to apologize for a racist message he sent in the 1990s that recently came to light.

However, his ideas about the dangers posed by artificial intelligence remain highly influential and have inspired both billionaire Elon Musk, CEO of Tesla and SpaceX, and astrophysicist Stephen Hawking, who died in 2018.

Some experts dismiss the danger of the most popular Terminator among the general public.

Image of a red-eyed robot with finishedsent from the future by artificial intelligence to destroy all human resistance, marked the collective unconscious in a special way.

But according to experts from the Stop the Killer Robots campaign, autonomous weapons will not be imposed in this way in the coming years, according to a report published in 2021.

“Artificial intelligence won’t make machines want to kill people,” robotics expert Kerstin Doutenhan of the University of Waterloo in Canada told AFP when contacted.

“Robots are not evil,” she said, acknowledging that their creators can program them for harm.

A less discussed danger of AI

A less obvious scenario is that artificial intelligence will be used to create toxins or new viruses in order to spread them around the world.

Researchers from a group that uses artificial intelligence to discover new drugs conducted an experiment in which they modified artificial intelligence to search for harmful molecules. According to an article published in the journal Nature Machine Intelligence, they managed to create 40,000 potentially toxic substances in less than six hours.

With these technologies, someone could find a way to spread a “poison” like anthrax more quickly, said Joanna Bryson, an AI expert at the Herty School in Berlin.

“But it will not be an existential threat, just a terrible weapon,” she added to AFP.

Artificial intelligence can make homo sapiens an obsolete species

In apocalypse movies, disaster happens suddenly and everywhere at the same time. But what if humanity gradually disappeared, replaced by machines?

“In the worst case scenario, our species could become extinct without descendants,” predicts philosopher Hugh Price in a promotional video for Cambridge University’s Center for the Study of Existential Risks.

However, there are “less bleak possibilities” in which humans augmented with advanced technology can survive. “Then our purely biological species will disappear,” continued the same researcher.

In 2014, Stephen Hawking told the BBC that humans might one day no longer be able to compete with machines, and that this moment would be “the wake-up call of the human race”.

Geoffrey Hinton, a researcher who is trying to create machines that resemble the human brain, recently expressed for the Google Group in similar terms a kind of “superintelligence” that would be superior to that of a human.

He recently told the American television station PBS that perhaps “humanity is only a passing phase in the evolution of intelligence.”

PHOTO article: Horacio Selva / Dreamstime.