Sam Altman says cognitive atrophy is one of the most overlooked risks of the AI era

Sam Altman says 1 of AI's underappreciated risks is that group whitethorn fto the exertion do truthful overmuch reasoning for them that their ain intelligence muscles statesman to weaken.

"One that I don't deliberation gets overmuch attraction is really are we going to debar cognitive atrophy?" the OpenAI CEO said connected an section of the "Invest Like the Best" podcast that aired connected Tuesday, erstwhile asked astir the astir important unresolved questions facing AI.

"How are we going to usage these devices and make judge that we are stretching our brains much and much and continuing to understand the worldly that really matters?" Altman said.

His comments travel arsenic a increasing number of researchers person warned that while the exertion could make workers faster, it could besides weaken their underlying occupation and cognitive skills.

"There's ne'er been a instrumentality for illustration this that has offered america the opportunity to fto reasoning beryllium done for us," Nataliya Kosmyna, an MIT interrogator whose preliminary study recovered that group utilizing generative AI to constitute essays performed worse complete clip than participants utilizing Google aliases nary aid, told Business Insider earlier this month.

Vivienne Ming, a theoretical neuroscientist and AI researcher, has described the consequence of group abandoning cognitive activity because AI tin do it for them arsenic "chronic substitution." She told Business Insider past period that repeated substitution could weaken cognitive reserve, a cardinal defense against dementia.

Those concerns are particularly pressing arsenic AI devices move beyond answering questions and generating matter to taking connected progressively complex work, including archive review, research, analysis, customer support, and coding.

Altman isn't the only tech leader to person raised concerns astir AI's effect connected cognition.

Last year, Arthur Mensch, the CEO of Paris-based Mistral AI, told The Times that AI's biggest consequence whitethorn beryllium "deskilling" workers and making group progressively lazier arsenic they progressively trust connected the exertion to hunt for information.

Still, during the podcast interview, Altman struck a much affirmative reside astir what lies ahead.

He described a "pretty magical" early wherever a individual AI could beryllium ever on, reference documents, listening to meetings, recalling emails from weeks earlier, and utilizing spare computing powerfulness overnight to make ideas and propose what they should do next.

However, he said that moreover powerful models whitethorn struggle to replicate the qualities that make humans distinctive.

"The world whitethorn request for illustration a caller benignant of connection for the benignant of judgement that group are very bully at, that AI seems to really profoundly struggle with," he said.