Agencia EFE reported that Geoffrey Hinton, the 2024 Nobel laureate in physics, urged AI developers to build lasting concern for people into systems that may eventually rewrite their own code.
Hinton made the case at Starmus VIII in Tenerife on Oct. 20, EFE reported. He compared the relationship between people and a more capable AI system to a parent caring for a dependent child. The analogy raised whether a system with greater abilities could still be inclined to protect those more vulnerable than itself.
EFE reported that Hinton questioned whether developers could give AI a protective disposition that would persist even if a system could modify its own code. He also said it was plausible that AI could exceed human performance across almost all intellectual tasks within the next decade.
Hinton criticized major technology companies for giving too much attention to making AI more capable and too little to safety, EFE reported. He also argued that AI’s potential benefits were too great to simply abandon the technology.
In an interview with The Telegraph at Starmus, Hinton said: “So we should make them so they really care about us. And if we don’t do that, we’re toast.”
The Starmus program listed Hinton for an Oct. 20 panel, “Can Machines Understand the World?,” and a talk, “Making Alien Beings.” The festival ran Oct. 17-22 in Tenerife and La Palma. The Royal Swedish Academy of Sciences awarded Hinton and fellow Nobel laureate John J. Hopfield the 2024 Nobel Prize in Physics for discoveries and inventions that enable machine learning with artificial neural networks.
