Tom Merritt, tech journalist at DTS (and friends) discussed a Wall Street Journal article about AI tools being used by teachers to cutdown on time spent grading student papers and essays. Having retired from the classroom just as my colleagues were scrambling to deal with students employing chatbots to create their papers, my mind immediately bounced to a viral video that popped up some time ago of one ChatBot talking to another ChatBot:

So, if a student uses ChatGPT to do their essay and an educator uses an AI tool to provide feedback and grade the work, what are we really doing here? Is anyone doing their job? Isn’t the point of the essay/assignment to show student learning? Using an AI just shows that the student knows how to use a tech tool to create a facsimile of what the student thinks the teacher wants, but doesn’t show student learning on the actual subject in question.1  I get that all the non-educators are excited at the possibility of how this technology can relieve educators of the burden of having to provide timely/meaningful feedback to student work. You know what would be better than employing AI? Reduce class sizes so that humans can do meaningful work with other humans. Has anyone considered that option? 

I get it, it costs a lot to fully employ professional educators. The human is the most expensive resource in the classroom. And they do burn out after only a few years/decades. AI won’t tire out or require breaks or healthcare or affirmation. But, until we have actual GAI, what are we really doing? It feels like we’re spending billions to fund a possible AI solution when we’re unwilling to fund basic education. That’s just sad. 

In my 28-years teaching I employed and piloted every possible kind of technology to help lift some of the burden of being a classroom teacher. I used a business oriented contact management system (CMS/phone book), and copied all of my student’s contact info on to my laptop, so that I had the threat of instant access to parental contact info, should the need arise. Because the CMS was highly adaptable, I created “comment” fields where I could pre-compose the comments that would go on student report cards, that I would print out onto Avery labels, so that I could attach the “label”comments to the printed report cards. This way I could re-use comments that applied to students with similar situations, instead of writing everything out long hand, one student at a time. Using this method I could knock out my report cards in two days max, whereas the veteran teachers around me carried around their stacks of report cards, filling them out by hand for at least a week straight if not longer. 

When I was teaching at DeMille Middle School (LBUSD) in 2005, I got a chance to pilot a Learning Management System (LMS) called SchoolLoop where I could post assignments to a website and have students login to the website, access the instructions and in many cased do their work on the website. Because I was teaching in a computer lab environment, where every student was sitting in front of their own classroom computer, this was perfect. I had the class session warm-ups ready for them once they sat down and that’s what they did while I took attendance and did the other opening duties, etc. Then when the pilot program ended (and they took away our access to the service), I continued posting assignments using my iCal application (we were a Mac lab), and posted the instructions in the “Notes” field for each day and each period. 

I even bought into the “flipped classroom” philosophy, where if you pre-record the lesson to a video and give students access to the video, to pre-load instruction, then class time could be better used answering questions and working through issues with actual work, instead of droning on or lecturing. I liked the idea because, as a learner, if I wasn’t getting something, I could listen to the same section of the instructions multiple times and pause things as I was following the instructions. It worked for me… but most students weren’t as motivated as I was and it was more often that, after viewing the video, they’d raise their hands and ask, “What are we doing?” (Just like what they do after you spend several minutes giving them instructions live…). I learned that doing the videos was a convenience for me, but was only as “good” as the students’ willingness to cooperate and many of them just didn’t want to. Turns out, I was using a technological solution for what was, at its core, a sociological/community problem. It worked at DeMille in 2005, but that might have been helped by the novelty of getting instruction via computers (instead of teacher lectures). And by 2022 they couldn’t be bothered to even bring their Chromebooks to class. 

I’m all for lightening the load for teachers as much as possible, but maybe the AI should be used to do the parts that teachers really hate, like tracking down parents and guardians to communicate unacceptable student behavior or student lack of effort to do the work. Or they could employ the AI to automatically flag student work that was generated by other AI. There are fundamental things not working in the classroom today, like student accountability, missing administrative support, politicizing curriculum, “standardizing” instruction to the point of meaninglessness. Adding AI to the mix isn’t as much of a solution as AI-cheerleaders would have us think.

You want to “fix” education? Reduce class sizes so that the professional humans have the time to work with the learning humans. We’ve know for decades that “drill and kill,” which computer technologies are good at, only work for non-creative, non-solution rote-memorization, like how to install part A on to produce Z. It’s great for tasks that need to be done the same way over and over and over again. It’s great for making humans do the jobs that robots are taking over. It’s great for making human drones, but not thinking, problem-solving citizens. 

It would be wonderful if actual Artificial Intelligence could help a human understand why their sentence didn’t make sense or how they missed the nuance they were trying to craft. But it’s much more likely that it’ll be one AI critiquing the work of another AI with no human in the loop. There’s a much more fundamental (and addressable) problem in the classroom, beginning with thinking that technology can create individualized boutique education when educators are only allotted the resources and time for factory-floor one-size fits all education. It’s not working! Besides, we know you’re doing this or advocating for this solution so that your startup can make billions and not to help educators. Someone at some point probably thought that this would be good for educators, but that’s not what’s driving the big train now. Sorry, give us smaller class sizes, time to work with students and their parents, help us build communities of learners and learning (remember that concept!), and how about giving us a little respect for taking on the challenge of teaching the next generation how to be decent, creative, productive problem-solvers. 

2002-06-04 twain computer lab joe
2002-06-04 twain computer lab joe

Sources:

Tags: AI in education, AI on AI action, not a tech problem, tech in the classroom, what teachers really want

Creative Commons License

JosephBruceBustillos.com (website) by Joseph Bruce Bustillos is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License

FOOTNOTES:
  1. Which is the same problem of high stakes testing showing that students know how to navigate the testing process and not entirely about actual understanding of the area of being tested on.[↩]