How do you make sure an artificial intelligence doesn’t turn on humanity? According to Elon Musk, one way to do it is to make it curious.
In an interview on Fox News Channel’s Tucker Carlson Tonight, Musk said he’s working on an AI platform called TruthGPT. The project will differ from competitors such as OpenAI’s ChatGPT and Google’s Bard, Musk claims, by caring about understanding the universe.
“I’m going to start something which I call ‘TruthGPT’ or a maximum truth-seeking AI that tries to understand the nature of the universe,” said Musk. The CEO of Tesla and Twitter thinks that an AI that cares about understanding the universe is “unlikely to annihilate humans” as we’re an “interesting part of the universe, hopefully.”
Musk didn’t share too many details about the project, and it’s unclear how, exactly, he plans to go about building this curious AI. He did say that he considers AI dangerous if mismanaged, with the “potential for civilizational destruction,” and called for some sort of government oversight over AI projects.
“It wouldn’t quite happen like Terminator,” he said.
Musk isn’t new to AI; he is one of the co-founders of OpenAI (the company that built the seemingly omnipresent AI chatbot, ChatGPT), having donated $100 million to the project when it was still purely a non-profit. He is, however, a busy man, with Tesla, SpaceX, Twitter, The Boring Company, and Neuralink to run, so it’s unclear how he plans to fit yet another project into his schedule.
Musk recently signed an open letter warning about the dangers of AI, urging that AI research companies halt training of AIs for at least six months.
