Unboxed Logo

Home / Consumer Tech

ChatGPT passes U.S. Medical Exam, Wharton MBA test

ChatGPT passes U.S. Medical Exam, Wharton MBA test
Advertisements
Microsoft on Monday announced a “multibillion-dollar investment” in OpenAI, the parent of the wildly popular AI app ChatGPT.

 

OpenAI’s ChatGPT is in the headlines again! This time the AI-based text generator has passed some of the toughest exams in the U.S., and that too with good grades!

 

Talking first about the United States Medical Licensing Exam (or USMLE), it is an exam consisting of three tests that are typically given by medical students to practice medicine and become licensed doctors in the States. ChatGPT passed all three exams without any specialized training or reinforcement, according to a research paper on medRxiv.

 

That’s not it. The AI bot demonstrated a high level of concordance and gave insights into how it came to the answers. “These results suggest that large language models may have the potential to assist with medical education, and potentially, clinical decision-making,” the research paper noted.

 

In another interesting instance, ChatGPT also managed to pass the exam of a typical MBA core course – Operations Management – offered by one of the finest B-schools in the world, the University of Pennsylvania’s Wharton School of Business. But that’s not it!

 

The ChatGPT-3 powered chatbot scored B to B- grade in the exam. A paper by University of Pennsylvania’s Professor Christian Terwiesch summarised ChatGPT’s “academic performance” in four points. First, the chatbot does an amazing job at basic operations management and process analysis questions. It did not just correctly solve the questions but also provided “excellent” explanations.

 

However, in the second point, the Professor highlighted that the chatbot is weak in math (a subject that typically terrifies most), and makes mistakes in relatively simple calculations at the level of 6th-grade math. Third, the chatbot could not handle more advanced process analysis questions. One thing to note here is that the learning of OpenAI’s chatbot is currently limited to the dataset available till 2021.

 

And finally, the Professor noted in the paper that ChatGPT is “remarkably good” at modifying and correcting answers in response to human hints.

 

Since its release in November last year, ChatGPT has been a center of excitement for many due to its human-like response. It is a “conversational” AI app that answers “followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.”

 

Elsewhere, the American multinational tech giant Microsoft on Monday announced a “multibillion-dollar investment” in OpenAI, the parent of the wildly popular AI app ChatGPT. While the actual amount and terms of the deal are still unclear, some reports suggest that Microsoft will invest $10 billion in OpenAI.

Advertisements

Advertisements