I see lots of teachers use it to generate questionaires / labs / worksheets / etc - traditionally time consuming and frustrating tasks. Its fast, but it will also often have poor questions with 'correct answers' that are either partially incorrect (because the generated question has ambiguous or generalized terms that can be interpreted in more than one way) - or entirely incorrect, for whatever reason. It will also often pose the same question in slightly different wording or scenarios, even within a very short set of questions like a six-question short-answer lab. This is even with supposedly 'top tier' paid LLMs like Claude.
Teachers often catch some of the mistakes in review, but I've not seen a single instance of them catching them all - which leads to students getting incorrect or misleading information. At the same time, teaching loads and administrative burdens placed upon them keep increasing, making them time-poor and stressed, so they will continue reaching for LLMs to save time. Couple this with research showing regular LLM use reduces the user's cognitive reasoning and deduction abilities.. They'll just catch fewer mistakes as time goes on.
This is not at all to pick on teaching, just to share my anecdote and agree that AI is churning out slop in every industry, as far as the eye can see.
I see lots of teachers use it to generate questionaires / labs / worksheets / etc - traditionally time consuming and frustrating tasks. Its fast, but it will also often have poor questions with 'correct answers' that are either partially incorrect (because the generated question has ambiguous or generalized terms that can be interpreted in more than one way) - or entirely incorrect, for whatever reason. It will also often pose the same question in slightly different wording or scenarios, even within a very short set of questions like a six-question short-answer lab. This is even with supposedly 'top tier' paid LLMs like Claude.
Teachers often catch some of the mistakes in review, but I've not seen a single instance of them catching them all - which leads to students getting incorrect or misleading information. At the same time, teaching loads and administrative burdens placed upon them keep increasing, making them time-poor and stressed, so they will continue reaching for LLMs to save time. Couple this with research showing regular LLM use reduces the user's cognitive reasoning and deduction abilities.. They'll just catch fewer mistakes as time goes on.
This is not at all to pick on teaching, just to share my anecdote and agree that AI is churning out slop in every industry, as far as the eye can see.