ChatGPT4 used the correct theorem on a question but erred in handling inequalities.
I uploaded this GMAT™ Official Starter Kit quant question to ChatGPT and requested a solution.
ChatGPT's response, however, was incorrect.
Specifically, ChatGPT correctly applied the theorem but mistakenly stated that 3 + 8 was greater than 11, rather than equal to it.
When I pointed out the error, ChatGPT admitted the mistake and provided the correct answer.
Why can’t ChatGPT provide the correct answer initially? In an exam setting, it doesn't inspire confidence if ChatGPT stumbles on simple tasks like comparing two integers.
Relying solely on AI for exam questions is not yet fully reliable without significant human supervision yet.
Our exam proxy experts remain indispensable.
For now.
