(I’m republishing my best posts from the first half of the year. You can see the entire list of them here)

 

geralt / Pixabay

 

Comparing the quality of human and ChatGPT feedback of students’ writing is a new study that seems very well done (plus, it’s not behind a paywall).

I particularly appreciated that they had an equal number of ELLs and non-ELLs as part of the study.

They found – to no one’s surprise – that human feedback to student writing was more accurate than AI feedback from ChatGPT 3.5.  However, there wasn’t that big of a difference between the two.

Of course, two critical elements were not included in the study, and which are critical in any kind of writing instruction.

One is did students feel motivated to actually implement the feedback that was given?  I suspect that in a real classroom many students are going to feel much more energized to do so if they know a human read it and is giving them their comments.

Secondly, how motivated will students be to write again?  Everyone, I believe, wants to have an “authentic audience” for their work, and I’m not sure that a chatbot will cut it as one.

From what I hear, more and more higher-ed instructors are using AI for feedback to student writing.  I guess with the numbers some of them have in their classes, I can understand why they might give it a try.

But I think it’s pretty clear that our K-12 students deserve better.