Nobel Prize in Physics in 2024, Briton Geoffrey Hinton does not rule out the possibility that artificial intelligence (AI) could destroy the world in less than three decades. This specialist, considered to be one of the fathers of artificial intelligence, nevertheless specifies that this would have between 10 and 20% risk of occurring.
The fear of out-of-control AI
The debate around potential danger that AI poses to humans is not new, but continues to make news. It must be said that in the past, certain inventions have already delivered quite worrying content, notably the ChaosGPT initiative in 2023. What if AI could really destroy humanity? According to the British Geoffrey Hinton, this would have between 10 and 20% chance of occurring by 2050.You should know that this man recently interviewed by BBC Radio 4 is considered one of the pioneers in the field of artificial intelligence. Let us recall in passing that this computer scientist and researcher specializing in this field won the Turing Prize in 2018 as well as the Nobel Prize in Physics in 2024.
Geoffrey Hinton therefore reaffirms his position on the subject. Indeed, he had already shared similar comments on the social network X in 2023 where he affirmed that AI had one in ten chance of destroying humanity in less than thirty years in the absence of regulations or in the event of poor management. If the specialist believes that artificial intelligence is capable of leading humanity to extinction, hope is still present. Geoffrey Hinton thus once again underlines the need to establish rules to ensure that this technology will not escape human control.
A highly anticipated AI summit in Paris
“Facing a three-year-old, we have control. But with AI, we are the three-year-old child”recalls the researcher. Geoffrey Hinton fears that some large companies motivated by profit develop out of control AI. Therefore, the only solution would be to force these companies through government measures to further research in order to achieve a satisfactory level of safety. It must be said that for the moment, there is no national or international regulation to constrain players in the field, apart from the European AI Act in force since March 2024.
Obviously, Geoffrey Hinton is not the only one to fear a future controlled by AI. Yoshua Bengio, another pioneer in this field, had already stated that AI was the most dangerous invention in human history. Finally, the two researchers should, however, welcome the holding of the Summit for Action on Artificial Intelligence in Paris on February 10 and 11, 2025. This event will bring together many heads of state and will be a good opportunity to implement regulations on a global scale about this technology.
Belgium
