For a long time, humans have gotten better and better at math. With new theorems and postulates came more mathematical knowledge. But now, there is a clear threat to even the smartest human mathematicians: AI’s ability to solve math problems.
Despite starting out vastly inferior to human reasoning, AI has caught up quickly. It has learned in short amounts of time wisdom that took many centuries for humans to discover. In fact, AI was able to prove the Four-Color Theorem in 1976. The theorem, which was the first proven by AI, states that only four colors are needed to color a map such that no two regions that touch have the same color.
With the recent advent of Chat-GPT and the exponential increase in the computing power of engines, many in the math sector are starting to feel threatened. In fact, developers are saying that AI may exceed human reasoning capabilities very soon. Computer scientist named Christian Szegedy even predicts that AI may outsmart humans by 2026.
What does this mean for people in the field of mathematics? Well, many believe that the era of AI is going to usher in some big changes. One mathematician, Akshay Venkatesh, who works at the Institute for Advanced Study at Princeton, said that while he wasn’t opposed to the thoughtful use of technology to improve humans’ understanding of math, he did have concerns about the “mindfulness” with which people will use it.
AI has already helped humans in math. In 2016, a researcher and his collaborators used AI to solve a math problem, and the solution took up more than 200 terabytes of data. However, this colossal amount of math led some to believe that the brute force tactics that computers often use can’t even be considered math.
Some say that feeding machines hundreds of proofs and theorems is time-consuming, and AIs often don’t understand a simple proof that humans can grasp with basic common sense. But others believe that the problem is much deeper. While AIs can provide a mathematical answer, many times they cannot give any reasoning or logic to help humans understand it and can only use brute force to solve equations and problems.
Others, meanwhile, just want the topic to be discussed more. Columbia University professor Michael Harris thinks that more debate and discussion is needed. A “very lively conversation is going on [about AI] pretty much everywhere except mathematics,” he said.
Despite starting out vastly inferior to human reasoning, AI has caught up quickly. It has learned in short amounts of time wisdom that took many centuries for humans to discover. In fact, AI was able to prove the Four-Color Theorem in 1976. The theorem, which was the first proven by AI, states that only four colors are needed to color a map such that no two regions that touch have the same color.
With the recent advent of Chat-GPT and the exponential increase in the computing power of engines, many in the math sector are starting to feel threatened. In fact, developers are saying that AI may exceed human reasoning capabilities very soon. Computer scientist named Christian Szegedy even predicts that AI may outsmart humans by 2026.
What does this mean for people in the field of mathematics? Well, many believe that the era of AI is going to usher in some big changes. One mathematician, Akshay Venkatesh, who works at the Institute for Advanced Study at Princeton, said that while he wasn’t opposed to the thoughtful use of technology to improve humans’ understanding of math, he did have concerns about the “mindfulness” with which people will use it.
AI has already helped humans in math. In 2016, a researcher and his collaborators used AI to solve a math problem, and the solution took up more than 200 terabytes of data. However, this colossal amount of math led some to believe that the brute force tactics that computers often use can’t even be considered math.
Some say that feeding machines hundreds of proofs and theorems is time-consuming, and AIs often don’t understand a simple proof that humans can grasp with basic common sense. But others believe that the problem is much deeper. While AIs can provide a mathematical answer, many times they cannot give any reasoning or logic to help humans understand it and can only use brute force to solve equations and problems.
Others, meanwhile, just want the topic to be discussed more. Columbia University professor Michael Harris thinks that more debate and discussion is needed. A “very lively conversation is going on [about AI] pretty much everywhere except mathematics,” he said.