"AI will never be as good as the best experts in a field."

Post Reply
User avatar
Posts: 1113
Joined: Mon May 17, 2021 3:03 pm

"AI will never be as good as the best experts in a field."

Post by funkervogt »

Ethan Mollick, an entrepreneurship and innovation professor at the University of Pennsylvania's Wharton School, told NPR on Thursday that he now requires his students to use ChatGPT to help with their classwork.

"This is a tool that's useful," Mollick said during the NPR interview. "There's a lot of positives about it. That doesn't minimize the fact that cheating and negativity are there, but those have been there for a long time."

His new AI policy — which NPR reviewed — calls AI usage an "emerging skill." The policy also states that students must check ChatGPT's responses and will be held accountable for any inaccuracies that the bot spits out.

..."AI will never be as good as the best experts in a field," Mollick said. "We still need to teach people to be experts."
https://www.businessinsider.com/wharton ... ing-2023-1

I'm posting this to point out the article's flaws. First, Ethan Mollick is doubtless a brilliant man, but nothing in his background indicates he knows more than the average person about AI or its ultimate potential. His declaration that it "will never be as good as the best experts" is hence not worth paying attention to.

Second, even if we assume he is right about AI, the consequences of the human race will be monumental. Even if some kind of inherent technological limitation prevents AI from ever becoming "as good as the best experts in a field," that won't stop AI from being better than all the mid-level experts and non-expert people in that field. Even if the most talented 1% of experts remain smarter and more competent than AI, the bottom 99% still become unemployable and useless.

Third, I want to highlight how this article ends with a feel-good statement that AI still isn't perfect and probably won't surpass humans. I see this kind of thing often, and view it as a dumb tactic meant to make readers feel comforted. The fact that AI isn't perfect yet and isn't superior to humans in all domains yet is not the important takeaway. Instead, the article should have concluded that AI is rapidly improving, has now entered the human realm of competence in several important areas and even exceeded it in some, and has an unknown upper limit on its capabilities. Yes, AI could top out at being better than ONLY the bottom 99% of humans, or it could vastly exceed us in all areas.
Post Reply