▾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.
Wi-Fi 6: Faster, more stable, and more efficient internet for your home and office
Wi-Fi 6, the latest wireless network standard, brings a revolution in speed, capacity, and efficiency. With new technologies like OFDMA, MU-MIMO, and BSS Coloring, it can provide up to four times the throughput and serve more devices simultaneously. Find out how Wi-Fi 6 can help you at home and in the office.
Programming for kids: fun and a skill for the future
Programming is not just about writing codes, it is a path to developing logical thinking and creativity. Discover how you can introduce children to the world of programming in a fun way, from simple block languages to real coding.
the dark side of influencer marketing: how fake followers distort the reality of social media?
Fake followers are like digital doping - a quick path to popularity with devastating consequences. Discover with us the dark side of influencer marketing, where thousands of followers can be bought for a few hundred crowns. What impact does this phenomenon have on brands, society, and especially the younger generation?
OpenAI funds research on artificial intelligence morality
OpenAI is investing a million dollars into research at Duke University aimed at developing algorithms that predict human moral judgments. The research team previously created an AI system for decision-making in transplants. Current AI systems, however, operate solely on a statistical basis and lack true understanding of ethics. Furthermore, different AI systems uphold different philosophical stances on morality.
X launches a free version of the AI chatbot Grok
The AI chatbot Grok by xAI, previously available only to paying X users, is now open to the general public. The free version comes with a limited number of queries but still offers access to powerful features such as image generation and understanding. Does this mark a new era in the accessibility of artificial intelligence?
the online world in 2050: what awaits us in the digital future?
Do you want to know what awaits us in 2050? Forget flying cars, the real revolution will happen online. Artificial intelligence will advise us on what to wear, smart homes will take care of shopping, and in virtual reality, we will have coffee with a friend from across the world.