AI, aye yi yi...

So, I'm dipping my feet in the AI information regarding ethical use of the models and services in the school setting. I did an online modular training yesterday, and I signed up for a webinar hosted by the NEA and ISTE+ASCD that will be in a couple of weeks. I have no interest at all in using AI myself, but I put myself on a committee at school tasked with developing an AI use policy and methods of teaching about AI/ using AI for classrooms. I'm probably going to play the role of naysayer, but someone has to watch out for academic rigor.

I'm not a fan. I think, at the most innocuous, it is lazy to use AI when applied to writing and related tasks. It's inaccurate, although it's clearly being tweaked all the time to make it more accurate. That said, there was a punctuation error, an easy-to-fix one, staring me in the face on their presentation yesterday. (sigh) This, after they talked about using AI to review work for grammar and usage. 

Two things trouble me so far, given that I know teenaged students pretty darn well. One is that you can highlight text, and then prompt the program (service) to rephrase the text. Um... we call that cheating. And the other thing, which is equally concerning is that you can write a whole piece, then prompt the program (service) to find articles online that support your work. This takes all the scholarship out of writing, and runs very close to confirmation bias. No learning happens, either, which is the crux of the problem. 

Of course, the people in the webinar were all about "you are in charge, you have to check what it tells you" and so on-- but if you are a teenager, first, do you have enough knowledge about the topic you are writing about in order to check what AI spits out? No. That would be the whole purpose of finding scholarly work to read and digest before the paper is drafted. And second, there are precious few students who would check the work anyhow. The introduction of AI into the classroom will necessitate having frequent "milestone" checks on the students' work: do you have a topic? What is your thesis? Find your sources and develop an annotated bibliography to turn in. Write your "prethinking" draft (all you think you know) and have the teacher check it. Then, work with the source materials; do they amplify your points, or do they refute what you think? Draft again. Check with the teacher. Make sure to document your sources and cite them in the text. Draft again. Check with the teacher. Turn in an assessment draft. You'll likely have to go back to it one more time. 

That's how I do papers now; AI is not that new, and it is getting harder to check when kids use it-- except, there is a notable lack of individual voice, and the diction of sections of their work shows a hard shift. Students have plagiarized for ages, and AI is just one more way of getting things done without actually doing them. And then, of course, you can question the students: explain this part to me? This is where an oral presentation is quite useful, too. And I do that, too. 

All these steps are elements of good scholarship, but not every teacher employs them. The "one and done" papers are potentially garbage, and now, with the rampant use of AI, they are not even useful learning tools. That is, after all, why students need to do writing assignments: they need to read somewhat deeply, think, talk, write and rewrite, and talk again... this will help them discover what they know, and will add to their knowledge base. AI may have its uses (yay, you can make ad blurbs quickly, and you, too, can create click bait for online purposes), but in the classroom, if actual learning is the objective, AI circumvents the learning process. 

The policy we draft will exist in time and space, but it will not, I suspect, have enough muscle to make kids learn effectively. Giving our agency as human beings over to computers is, at the very least, unwise. More chillingly, I foresee a not-distant future of poor "scholarship" that is based in prior "poor scholarship," leading to very faulty understandings of subject matter-- and this could literally be life-or-death. 

What a brave new world, eh?

C