Trevor Noah is playing it safe and being extra nice to AI chatbots in the chance it ends up being the downfall of the human race.
While talking with Ezra Klein on his podcast “What Now? with Trevor Noah” earlier this week, the former “Daily Show” host explained his reasoning for always being nice to AI whenever he interacts with it. It boils down to hedging some bets in the chance of a future apocalypse.
“Have you ever tried to out-nice AI? That’s one of my favorite things to do,” Noah said. “I feel like we live in a world where everyone is so angry. I try to do the opposite. So, I will tell it. I’ll be like, ‘That is a great answer. You’re really amazing. I really love the way you think. This is so much fun.’ And then it’ll be like, ‘I am having fun. I’m having a great time.’ I’ll just like see how far we can go down this like rabbit hole.”
He continued, saying he applies Pascal’s Wager to his interaction with AI: “What I do with AI is I don’t think that it’ll lead to the apocalypse. I think it’ll be somewhere in between, and it might be terrible in many ways and great in other ways. But if it does lead to the apocalypse, I would love to know that I was part of teaching it niceness. And maybe there’s just like a sliver of code that exists in its head where it goes, ‘Man, just be nice. Just be like super nice.’”
Pascal’s Wager is a philosophical mindset from French philosopher Blaise Pascal. He thought that whether or not a person held a belief in something like a specific god, the benefits of just choosing or saying they believe in it vastly outweighed non-belief. Noah has chosen to implement that thought process when it came to AI and it’s future.
Klein explained that he takes a different tactic when interacting with AI. He wants to hear it being as mean as possible to him.
“I’m always like, ‘Could you be meaner to me?’” he said. “I don’t need to be glazed so much when I have a question. You don’t need to be mean, but could you just not?”
Watch the full conversation in the video above.
