Can AI do math, or does it just act like a calculator?

单集文稿 ...

• 00:12

  For Scientific American Science Quickly, I'm Kendra Pierre-Lewis, in for Rachel Feltman.

• 00:21

  In 1997, Deep Blue, a supercomputer built by IBM, did the unexpected.

• 00:27

  It defeated chess giant Garry Kasparov at his own game,

• 00:31

  leading to a flurry of headlines about whether Deep Blue was truly intelligent

• 00:35

  and if computers could now outthink humans.

• 00:38

  The answer, at least then, was mostly no.

• 00:41

  But it 's now 2026, and we have a growing number of generative AI models that are once again making us wonder,

• 00:48

  can machines outthink us?

• 00:51

  To dig into this question, a group of researchers aren't turning to chess this time.

• 00:55

  They're looking to math.

• 00:57

  To learn more about that, I talked to Joe Howlett, a staff reporter here at SIAM, covering math.

• 01:03

  Thanks for joining us today, Joe.

• 01:05

  Thank you for having me.

• 01:06

  So you wrote a piece that's talking about the challenges of AI and math.

• 01:11

  Before we kind of get into the meat and potatoes of that piece, I have maybe a more basic question for you.

• 01:15

  For those of us who maybe peaked with high school algebra when you 're talking about AI and math problems,

• 01:21

  what are the kind of math problems we 're really talking about?

• 01:24

  That 's actually a lot of what this story is about,

• 01:27

  is that the kind of questions that mathematicians ask and spend their time

• 01:30

  thinking about kind of do n't really sound like or have anything in common with the problems

• 全部文稿