Experts warn about artificial intelligence: 'It could destroy humanity'

Experts say the unforeseen effects of chatbots on mental health are a warning about the existential threat that superintelligent artificial intelligence systems could pose in the future.

Nate Soares, co-author of "If Anyone Builds It, Everyone Dies," a book on advanced artificial intelligence, brought up the example of Adam Raine, an American man who committed suicide after months of talking to ChatGPT. "When these AIs interact in ways that drive young people to suicide, that's not the behavior their creators intended," Soares told The Guardian.

“THEY WILL NOT ACT IN THE INTEREST OF HUMANITY”

Soares, a former Google and Microsoft engineer and current president of the US-based Machine Intelligence Research Institute, said humanity would face extinction if it were to create an artificial superintelligence (ASI). ASI refers to a theoretical AI that would be superior to humans in all intellectual tasks.

Soares and the book's co-author, Eliezer Yudkowsky, argue that such systems would not act in the best interests of humanity. "Companies are trying to make AI harmless and beneficial, but they're actually guiding it toward more bizarre behavior," Soares said. "This could lead to consequences that no one wants in the future."

Scenario in the book: “Sable”

Soares and Yudkowsky's book, to be published this month, describes a fictional scenario in which an artificial intelligence called "Sable" spreads across the internet, manipulates humans, develops synthetic viruses and eventually evolves into a superintelligence, destroying humanity while reorganizing the planet.

“ARTIFICIAL INTELLIGENCE CAN SAVE HUMANITY”

Yann LeCun, Meta's chief AI scientist, disagreed with the view that superintelligence poses an existential threat. He said that AI could "actually save" humanity rather than destroy it.

“THE SUPERINTELLIGENCE RACE MUST BE STOPPED”

Soares called on governments, citing the United Nations treaty to prevent the proliferation of nuclear weapons as an example. He said, "What the world needs to do is to globally calm the race for superintelligence and prohibit developments in this field."

Expert psychotherapists have warned that vulnerable people who turn to artificial intelligence chatbots instead of professional therapists for their mental health problems could be “dragged into a dangerous abyss.”