Elon Musk says there's still a decent chance AI could go bad
The xAI founder made the comments during Saudi Arabia's Future Investment Initiative
As he anticipates the growth of his supercomputer, Elon Musk said there’s a chance artificial intelligence could go bad by the end of the decade.
While AI is “most likely going to be great,” Musk said “there’s some chance, which could be, 10 to 20%, that it goes bad,” during an interview with Peter Diamandis at the Future Investment Initiative in Riyadh, Saudi Arabia on Tuesday.
Diamandis asked Musk about predictions the billionaire made at the Abundance Summit in March that AI is growing “as fast as 100 times per year,” and that by the end of the decade, AI could become “as capable as 8 billion humans.”
“I certainly feel comfortable saying that it’s getting 10 times better per year, which is, four years from now that would mean 10,000 times better, so maybe 100,000,” Musk said about the pace of AI. “I think it will be able to do anything that any human can do — possibly within the next year or two.”
Three years after that, Musk said, AI could get to the point of doing “what all humans can do, combined.”