The AI-powered chatbot ChatGPT from San Francisco-based OpenAI has passed multiple tests in its research release version. With assistance from human tutors from University of Minnesota courses, ChatGPT passes legal and medical tests.
The Chatbot succeeds on tests for law from University of Minnesota courses, portions of the US Medical Licensing exams, and a test from the Wharton School of Business at the University of Pennsylvania.
Despite passing these tests, the chatbot occasionally needed help from a human and did not always do so with flying colours.
Will Chat GPT3 Attend Wharton for an MBA? A study was released and published. Using a test that MBA students normally complete, researcher Christian Terwiesch assessed the performance of the chatbot.
Terwiesch observed that ChatGPT occasionally produced “surprising blunders” in computations suited for middle school but performed “amazingly” when it came to essential questions regarding its subject. ChatGPT was occasionally able to alter its responses in response to cues, but it was still unable to respond to especially difficult questions.
Following the denial of rumours that the upcoming ChatGPT version would be substantially more advanced than the one that has been made available for public testing since December 1, 2022, Sam Altman, CEO of OpenAI, made the announcement.
Academics and test administrators have issued warnings about the dangers of students using ChatGPT to submit plagiarised essays or cheat on examinations, and corporations are working on a number of ways to spot the use of AI-based technology in submitted work.
