- The nominal Godfather of AI said the technology could develop a language humans can’t understand.
- As of now, AI thinks in English, meaning developers can track its thoughts — but that could change.
- His warning comes as the White House proposes limiting AI regulation.
Linguists may have their work cut out for them.
Geoffrey Hinton, the so-called “Godfather of AI,” warned that there could come a point when humans can’t understand what AI is thinking or planning to do. As of now, AI does “chain of thought” reasoning in English, meaning developers can track what the technology is thinking, Hinton explained on an episode of the “One Decision” podcast that aired July 24.
“Now it gets more scary if they develop their own internal languages for talking to each other,” he said, adding that AI has already demonstrated it can think “terrible” thoughts.
“I wouldn’t be surprised if they developed their own language for thinking, and we have no idea what they’re thinking,” Hinton said. He said that most experts suspect AI will become smarter than humans at some point, and it’s possible “we won’t understand what it’s doing.”
Hinton, who spent more than a decade at Google, is an outspoken about the potential dangers of AI and has said that most tech leaders publicly downplay the risks, which he thinks include mass job displacement. The only hope in making sure AI does not turn against humans, Hinton said on the podcast episode, is if “we can figure out a way to make them guaranteed benevolent.”
Tech companies are racing to get ahead in the AI race, offering gargantuan salaries to top talent. On July 23, the White House released an “AI Action Plan” that proposes limiting AI-related funding to states with “burdensome” regulations. It also asked for faster development on AI data centers.
Read the full article here
