Don’t believe everything a Large Language Model like #ChatGPT tells you. I have been experimenting with examples for my #moodle #aitext question type.

It accepts the answer X=1 rather than the correct X = 2.

The prompt was “Given the task of simplifying 3x +2 = 8. Is the following a correct solution, give reasons, indicate if the answer is wrong."

I am confident a more refined prompt would spot the error, but it shows how caution is necessary.
#edtech #edutoots #ai #aotext #chatgpt