Back in the 2010s many people thought that by 2025 we would have flying cars and robot servants. Instead of living up to those expectations, the 2020s have given us AI. Currently, the trends of technology have led us to artificial intelligence chatbots with powers of generating images, answering questions, and writing essays. But are they advancing to the point where they are credible?
One of the most commonly known chatbots, ChatGPT, can also do math based on what they have been taught. The process of teaching AI is either using the information already plugged in by the creators, or searching through online resources to supplement their knowledge base. Unfortunately, chatbots aren’t as capable in mathematics as they are in writing.
“The A.I. chatbots have difficulty with math because they were never designed to do it,” said Kristian Hammond, a computer science professor and AI researcher at Northwestern University. It seems as if AI was created more for liberal arts majors than mathletes.
Khan Academy, a widely known education nonprofit that is experimenting with AI, is figuring out why it isn’t as accurate in this category and how to make it better. A few months ago, their AI-based tutor named Khanmingo went through more reworks. Now, it sends numerical problems to a calculator instead of asking AI to slowly do the math.
Instead of using operations that only performed well on tasks such as generating images, Kristen DiCerbo, chief learning officer of Khan Academy, believes that using tools that do math will be better, providing more accurate answers and explanations. “We’re actually using tools that are meant to do math,” said Dr. DiCerbo.
Occasional slip-ups by the bot can transform into learning opportunities for students and a teaching one for teachers. When chatbots get different answers, students can compare their answers to the bot’s, seeing how each side got to their solution.
Students now know to not trust AI for everything. Maybe it can be useful and versatile in generating things, but right now math, AI, and accuracy currently do not belong in the same sentence.
One of the most commonly known chatbots, ChatGPT, can also do math based on what they have been taught. The process of teaching AI is either using the information already plugged in by the creators, or searching through online resources to supplement their knowledge base. Unfortunately, chatbots aren’t as capable in mathematics as they are in writing.
“The A.I. chatbots have difficulty with math because they were never designed to do it,” said Kristian Hammond, a computer science professor and AI researcher at Northwestern University. It seems as if AI was created more for liberal arts majors than mathletes.
Khan Academy, a widely known education nonprofit that is experimenting with AI, is figuring out why it isn’t as accurate in this category and how to make it better. A few months ago, their AI-based tutor named Khanmingo went through more reworks. Now, it sends numerical problems to a calculator instead of asking AI to slowly do the math.
Instead of using operations that only performed well on tasks such as generating images, Kristen DiCerbo, chief learning officer of Khan Academy, believes that using tools that do math will be better, providing more accurate answers and explanations. “We’re actually using tools that are meant to do math,” said Dr. DiCerbo.
Occasional slip-ups by the bot can transform into learning opportunities for students and a teaching one for teachers. When chatbots get different answers, students can compare their answers to the bot’s, seeing how each side got to their solution.
Students now know to not trust AI for everything. Maybe it can be useful and versatile in generating things, but right now math, AI, and accuracy currently do not belong in the same sentence.