Just got back from a few much needed days away so it's time to get to those chatGPT posts.

Last time, I wrote on cheating in general. That seems to be one of the hot points for chatGPT, in fact, the NYC DOE just last week decided to ban chatGPT outright, presumably for that reason. We'll talk about the student side, both negative and postitive today and the teacher side next time.

To start, while on vacation I saw this tweet:

Any educator who is saying “ChatGPT won’t affect MY classroom” better get humble really quickly.

As a friend said, “This is like when my AP called me into his office & asked ‘Have you seen this thing called Google?’”

It’s going to change the way we think about writing forever.

— Chris Lehmann (@chrislehmann) January 7, 2023

Sure, it's going to change our classrooms but it's not like this is the first time. We went from nothing to calculators to graphing calculators to programs like Wolfram Alpha. I've been telling my students for decades that Emacs Calc mode could not only solve all their math assignments symbolically but it could also typeset them directly in TeX so their homeworks would look professional.

On the non-math side we've added word processors to spell checkers to grammar aides not to mention collaborative environments.

And we haven't even mentioned search engines, wikipedia, youtube and the scores of other resources available today that didn't exist years ago.

New technologies come in to play over and over again and teachers adapt. Unfortunately, instead of giving teachers the resources they need to adapt efficiently, districts normally opt for a hammer and then blame teachers regardless.

So, now we have a new tool - will students use it to cheat? Sure - some will but this is nothing new. I'd like to believe that the majority of students will use the new tool productively but to do so, many will need guidance.

Should the NYC DOE have banned chatGPT? I'd say no. Well, actually, I'd say yes but with a caveat and we'll get to that below.

I've been playing with chatGPT a bit now and it certainly seems to be able to generate student quality essays on a variety of topics. It 100% can solve introductory CS programming problems, frequently just by pasting the teachers prompt in the chat box.

This can, of course, lead to cheating but shutting down the technology in school won't change a thing. As usual, draconian measures will do little to deter bad behavior and will merely make things more difficult for kids who want to do things the right way.

Let me use GitHub as an example. I've had many professors tell me that they'll have students use GitHub but only if repos are private so as to prevent cheating. Right - making a repo private is really going to prevent some kid from sharing their code. Sure, it'll prevent a random student from stumbling on their work but it's going to do nothing to prevent a bad actor from acting badly.

At the same time, it sows distrust and prevents students who want to productively share and help each other (with proper attribution) from doing so.

Banning a tool in school only serves to limit students who want to do things the right way anyway. Cheating happened prior to chatGPT and prior to older technologies. Some students have always cut corners either by crossing the line with friends or even paying for essays online. Now with chatGPT, I can type in "write a program to do…" rather than search StackOverflow or ask a friend or pay a service but it doesn't change the fact that no matter what, some students will cheat. On the other hand, if you cultivate a classroom where support and sharing is encouraged, don't overwork students, provide the needed resources and allow for failure and growth, you'll minimize the problems. Of course I just described a panacea that can rarely be created in a public school given large class sizes, emphasis on standardized testing, and all the baggage dumped on teachers.

In a programming class, a solution could be to allow chatGPT but to include lab time in the class. During that time, the teacher can observe students and really get to know what they know - their strengths and weaknesses. I always was able to include lab time at Stuy and before that at Seward Park and by the end of the term, I knew my students, what they could do, and where they had challenges. To be honest, I could assign letter grades from that info alone and the grades would pretty much match the test results.

This can be done but it's getting more and more challenging given the weight placed upon teachers.

Given this, in my class, I don't want to ban chatGPT but want to take the time to figure out how to use it productively.

Why then did I say earlier that districts should ban it? Because at the end of the day, schools and teachers should set these policies but these policies need to be supported by the district. The district has to have the principal's back and the principal, their teachers.

A district should say "here's the boilerplate policy to ban chatGPT so if in your school or class it's inappropriate, you can ban it and we've got your back. On the other hand, a principle can implement a looser policy and that's fine and we encourage schools to implement policies where chatGPT can be used effectively as a learning tool" and then provide some boilerplate language for that.

The district has to have the hardest line because in some cases, that hard line will be either appropriate or necessary and then schools can loosen the restrictions as appropriate.

So, that's my take.

Next time - chatGPT as a teacher aide.