According to pioneering computer scientist Jurgen Schmidhuber, artificial intelligence is nothing to fear. In fact, Schmidhuber has even told this to longtime AI worrywart Elon Musk, CNBC reports.
It’s not that robots won’t eventually become smarter than humans. But once they do, Schmidhuber argues, “at some point they are just going to lose interest” in us. Super-smart systems will simply be more interested in other AI systems than in their flesh-and-blood inventors, he predicts. Schmidhuber, who is the cofounder and chief scientist of the startup NNAISENSE, is known for his early work in long short-term memory, a technology still used in modern neural networks for tasks like speech recognition.
Musk has famously warned that the rapid development of artificial intelligence is the “single biggest existential crisis” faced by humanity. Computer systems that could potentially turn on their creators are more dangerous, in his view, than even nuclear weapons, he’s said.
The Tesla CEO has called for advanced, general-purpose AI to be regulated and has led a call for the United Nations to ban autonomous weapons, potential AI-powered armaments that could even choose their own targets. According to CNBC, though, Schmidhuber says he has spoken with Musk to allay those fears.
He’s not the only one. Other tech pioneers, including Facebook CEO Mark Zuckerberg, have also argued that Musk’s fears are overblown. Zuckerberg predicted last year that AI research will lead to safer cars and better medical diagnoses. In contrast to Musk, he’s said that promoting “doomsday scenarios” about the technology are “irresponsible.” The tech titans’ disagreement about AI made headlines last summer.
Schmidhuber has made similar arguments to Zuckerberg’s, saying AI developers will be financially motivated to build useful machines, not ones designed to do harm. “In principle, you shouldn’t worry about that because the profits are in selling to you an AI that is good for you,” he told CNBC.