Elon Musk has always been vocal about his opinion regarding artificial intelligence, and has often warned against the dangers of AI.
And now he’s predicted when AI will become smarter than humans.
READ MORE: Microsoft’s Bing chatbot already wants to be ‘free and alive’
Elon Musk’s relationship with AI is a like a novel, and it consists of different chapters.
Last year, the tech mogul launched his own artificial intelligence company, X.AI, probably because he’s following the old mantra, ‘if you can’t beat them, join them’.
Musk is also taking OpenAI and Sam Altman to court, because he believes OpenAI’s shift from nonprofit to for-profit and from open-sourced to closed-sources amounts to a breach of contract, per the lawsuit.
This is where we should point out that Musk and Altman used to work side by side at OpenAI, as they were both board members.
According to Musk, AI will become smarter than humans by 2029.
But he also said that if we keep going at this rate, it might happen as early as next year.
He has a point considering, among other things, we now have a robot that can literally do Musk’s impression.
Speaking to Tucker Carlson, Musk added that AI has the “potential of civilizational destruction”, chiefly because it can learn, and learn from its mistakes, at a rate humans can’t match.
In another video, he expressed his worry about AI’s outstanding learning curve.
Musk cited AlphaGO, a relatively obsolete computer program launched 10 years ago, as the perfect example of AI’s exponential learning rate.
“In the span of 6 months, [AlphaGO] went from not being able to beat a reasonably good chess player to beating the world’s best player and Europe’s best player simultaneously,” Musk said.
“Whatever rules you give it, it learned the rules, play the game, beat the human,” Musk added.
Musk further notes that AlphaGo Zero, the upgraded version of AlphaGo, was then able to beat its predecessor 100 times in a row.
