Wharton School 'Operations Management' Course at 'B' Level
Sometimes Makes Calculation Errors at 6th Grade Elementary Level
[Asia Economy Reporter Kim Hyunjung] Artificial intelligence (AI) scored above average on a required course exam at a top U.S. business school.
According to the website of the Mack Institute for Innovation Management at the Wharton School of the University of Pennsylvania (UPenn) on the 21st (local time), Professor Christian Terwiesch of the university published a white paper titled "Can ChatGPT3 earn a Wharton MBA?" on the 17th.
ChatGPT3 is the name of a conversational AI model developed by OpenAI, an AI-based startup. GPT stands for "Generative Pre-trained Transformer," a language model that uses deep learning to generate text similar to that used by humans.
Professor Terwiesch had ChatGPT3 take the final exam for "Operations Management," a required course in the Wharton MBA program. He evaluated ChatGPT3's exam results as likely earning a "B" or "B-" grade. While not an outstanding level, it appears to have scored better than at least some students. Founded in 1881, Wharton is the first business school in the U.S. and ranks among the top two prestigious business schools in the country for its MBA program.
In particular, ChatGPT3 reportedly demonstrated remarkable ability to automate some of the skills possessed by highly paid knowledge workers such as analysts, managers, and consultants. Professor Terwiesch stated, "ChatGPT3 solved basic operations management and process analysis problems very well. Not only did it get the answers right, but its explanations were excellent," adding, "Even when it failed to find the appropriate solution, if a human expert provided suitable hints, the AI was able to correct itself."
However, ChatGPT3 showed surprisingly weak performance in mathematics. According to Professor Terwiesch, ChatGPT3 occasionally made mistakes in relatively simple calculations at about a 6th-grade elementary school level and was unable to handle more advanced analytical problems.
Regarding this exam, Professor Terwiesch commented, "These results have important implications for business school education," and expressed the opinion that "there is a need to change exam policies and curriculum design to focus on human-AI collaboration."
OpenAI, founded in 2015 with investments from Tesla CEO Elon Musk and others, introduced the deep learning-based AI model GPT3 in 2020. GPT3 is designed to provide the most appropriate answers to questions using up to 175 billion parameters. Tasks GPT3 can perform include solving various language-related problems, random writing, simple arithmetic, translation, simple web coding based on given sentences, and conversation. OpenAI released a beta version of ChatGPT based on GPT3.5 last month.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
