Everyone’s heard of A.I. Coded and made to be a better version of humans, rumors have even arisen that A.I. might one day take over the world! But the truth is, sometimes, this “perfect” version of humans can’t get math problems right. While A.I. seems to be relatively good at everything else, simple arithmetic problems can stop A.I. in its tracks.
A.I. is fine-tuned for determining possibilities. But math—there’s usually only one answer in math. And yet many A.I. chatbots can give wrong or multiple answers.
Kristian Hammond, a computer science professor and artificial intelligence researcher at Northwestern University explains that, “The A.I. chatbots have difficulty with math because they were never designed to do it.”
A.I. was invented and created in this modern era. It learns by studying large amounts of data, like history, writing, or science. Based on the information it processes, it gives answers to questions humans ask it.
Dr. Hammond said, “This technology does brilliant things, but it doesn’t do everything. Everybody wants the answer to A.I. to be one thing. That’s foolish.”
It was designed and programmed to work like a smarter, more efficient human mind, to help humans answer questions–math included.
Just like humans, these A.I. chatbots need math practice from textbooks, drills and standardized tests. Then these chatbots can use the methods to answer similar questions. Recently, ChatGPT got 89th percentile in the high school math SAT test.
Even though A.I. isn’t the best at doing math, many websites and companies have used it in their products.
For example, Grammarly and Khan Academy have used A.I. chatbots in their products, while Meta used A.I.-powered smart assistant softwares in Facebook, Instagram, and WhatsApp. These assistance chatbots can chat with you or give you tips on how to cook, read, or write. But sometimes A.I.’s math skills can be quite frustrating.
While A.I. excels at lots of things, math certainly isn’t one of them. You can be rest assured that A.I. is still a long, long way from taking over the world.
Image Credit by Leeloo The First
A.I. is fine-tuned for determining possibilities. But math—there’s usually only one answer in math. And yet many A.I. chatbots can give wrong or multiple answers.
Kristian Hammond, a computer science professor and artificial intelligence researcher at Northwestern University explains that, “The A.I. chatbots have difficulty with math because they were never designed to do it.”
A.I. was invented and created in this modern era. It learns by studying large amounts of data, like history, writing, or science. Based on the information it processes, it gives answers to questions humans ask it.
Dr. Hammond said, “This technology does brilliant things, but it doesn’t do everything. Everybody wants the answer to A.I. to be one thing. That’s foolish.”
It was designed and programmed to work like a smarter, more efficient human mind, to help humans answer questions–math included.
Just like humans, these A.I. chatbots need math practice from textbooks, drills and standardized tests. Then these chatbots can use the methods to answer similar questions. Recently, ChatGPT got 89th percentile in the high school math SAT test.
Even though A.I. isn’t the best at doing math, many websites and companies have used it in their products.
For example, Grammarly and Khan Academy have used A.I. chatbots in their products, while Meta used A.I.-powered smart assistant softwares in Facebook, Instagram, and WhatsApp. These assistance chatbots can chat with you or give you tips on how to cook, read, or write. But sometimes A.I.’s math skills can be quite frustrating.
While A.I. excels at lots of things, math certainly isn’t one of them. You can be rest assured that A.I. is still a long, long way from taking over the world.
Image Credit by Leeloo The First