New research conducted by a professor at the Wharton School of the University of Pennsylvania found that the AI-powered chatbot GPT-3 was able to pass the final exam for the school’s Master of Business Administration (MBA) program.
Prof. Christian Terwiesch, author of the research work «Would Chat GPT3 get an MBA from Wharton? A prediction based on its performance in the operations management course,» he said the bot scored between B- and B on the exam. The bot’s score shows its «remarkable ability to automate some of the skills of highly paid knowledge workers in general and specifically of knowledge workers in roles held by MBA graduates, including analysts, managers, and consultants.»
The bot did an «incredible job addressing basic operations management and process analysis questions, including those based on case studies,» Terwiesch wrote in the article, which was published Jan. 17, adding that the «Bot explanations are excellent.» The bot is also «remarkably good at modifying its responses in response to human suggestions,» she concluded.
Terwiesch’s findings come as educators grow increasingly concerned that such chatbots could inspire cheating. Although chatbots are not a new technology, ChatGPT exploded on social media in late 2022. Earlier this month, the New York City Department of Education announced a ban on ChatGPT on its school devices and networks.
Much of the discussion centers around ChatGPT’s conversational chat style and coherent, thematic response style, making it difficult to distinguish from human responses.
Experts working in both artificial intelligence and education have recognized that bots like ChatGPT could be detrimental to education in the future. But in recent interviews with NBC News, some educators and experts weren’t worried—yet.
Terwiesch did not immediately respond to a request for further comment. A spokesperson for the artificial intelligence startup OpenAI, which created the bot, also did not immediately respond to a request for comment on Monday.
He GPT-3 The model used in the experiment appears to be a big brother to the more recent ChatGPT bot that has become a controversial topic among educators and those working in the AI field. ChatGPT, the latest version, «is refined from a GPT-3.5 series model,» according to OpenAI’s website.
While Chat GPT3’s results were impressive, Terwiesch noted that Chat GPT3 «sometimes makes surprising errors in relatively simple calculations at the sixth-grade math level.»
The current version of Chat GPT «is not capable of handling more advanced process analysis questions, even when based on fairly standard templates,» Terwiesch added. «This includes process flows with multiple products and problems with stochastic effects such as demand variability.»
Still, Terwiesch said ChatGPT3’s performance on the test has «important implications for business school education, including the need for exam policies, curriculum design focused on human-AI collaboration, opportunities to simulate real-world decision-making processes, the need to teach creative problem solving, improving teacher productivity, and more.”