▾Artificial intelligence writes poems but struggles with math. Why can't ChatGPT and other chatbots handle even basic arithmetic? We reveal the causes of AI's mathematical mistakes, from tokenization which breaks numbers into unintelligible fragments, to the statistical learning approach that fails in mathematics.
Artificial intelligence, including ChatGPT, can write poems, compose music, and translate texts. Yet, it often stumbles on simple mathematical tasks. Why can't a chatbot, that handles complex language tasks, deal with math at an elementary school level?
Tokenization: When numbers break into pieces
One of the key problems is tokenization. This process divides data into smaller parts, called tokens. Imagine it like assembling a puzzle, where words are broken down into syllables. The tokenizer, the AI model responsible for this process, does not understand the meaning of numbers.
It may happen that the number 380 is perceived as one token, while 381 is perceived as two (38 and 1). This disrupts the relationships between digits and complicates the calculation.
A statistical machine that falters with numbers
Another reason for ChatGPT's mathematical difficulties is its statistical nature. The chatbot learns based on a vast amount of examples and looks for patterns in them. For instance, it learns that the phrase "Dear Sir" is often followed by the phrase "we are reaching out to you".
However, this approach faces challenges in mathematics. ChatGPT can guess that the product of numbers ending in 2 will end in 4, but it cannot handle intermediate results. Simply put, the ChatGPT model tries to guess the result based on learned patterns instead of performing a precise calculation.
The challenge of multiplication
A study conducted by Yuntian Deng from the University of Waterloo showed that ChatGPT struggles with multiplying numbers greater than four digits. The reason is that any error in a calculation step shows up in the final result.
Imagine it as a domino effect – one error triggers a chain reaction, and the result is completely off. However, there is hope that ChatGPT will improve in the future. Deng and his colleagues also tested the o1 model from OpenAI, which is characterized by logical reasoning capabilities.
This model achieved significantly better results than the standard GPT-4o and was able to correctly solve multiplications of nine-digit numbers. The o1 model thinks through the problem step by step, allowing for more accurate results.
Security threats in your home: How to protect your router from hackers?
The router is the key to protecting your home network. That's why it is targeted by most hackers, who can not only access your sensitive data and files but also use it for further attacks. Find out how to protect yourself from them.
Anonymity on the internet: Is it truly possible to remain invisible?
Anonymity on the internet attracts many users, but is it really achievable? Discover the power of tools like VPN and Tor, which promise invisibility in the online world. Our guide will show you how to minimize your digital footprint and navigate the cyber space safely.
Google plans to power artificial intelligence with reactors
Google has partnered with Kairos Power and plans to power its data centers with small modular reactors. The aim is to secure 500 MW of zero-emission energy by 2030. The ambitious plan faces both technical and societal challenges.
AI in schools: How is it already changing our children's education?
Robots in schools? Don't worry, teachers won't lose their jobs. Artificial intelligence (AI) is indeed changing education, but according to experts, human creativity and critical thinking are still key. Find out how ChatGPT and other AI tools can assist both students and teachers.
Meta introduced revolutionary Orion glasses: The future of augmented reality is here
Meta introduced Orion, the world's most advanced AR glasses. They combine the appearance of regular glasses with augmented reality capabilities. With a holographic display and integrated AI, they open new possibilities for interaction with the digital world. In addition, they allow hands-free video calls, display messages, and provide real-time contextual information.
How to protect children on the internet? Try our tips
Children's safety on the internet is largely the responsibility of parents. Therefore, we have prepared a comprehensive guide to protecting children online. You'll learn how to communicate openly with them about risks, set boundaries, and protect their privacy.