“It feels like … cheating”

Ok, so as Will, Mike, Catrina and I were filming our superhero summary of learning this past week, I asked one of my students to pretend she was cheating via ChatGPT, and then I, ala ChatWoman, would obviously swoop in and deter her from this choice. When I asked her to do this, one of the other students popped in and said, jokingly, “Ohhhh  Chat GPT ohhhhh I wonder who is using this and why….”, as if to imply that she knew (or perhaps assumed) many students were using this to cheat. From there, the conversation turned to ChatGPT, and students asked when they were going to learn more about this. At the beginning of the semester I had mentioned it in the syllabus with regard to academic dishonesty, and then also mentioned that I would talk about how to use it to our benefit. To be honest, I had been putting off the lesson as I wasn’t quite sure how to frame it, and how much to talk about it. I don’t yet feel like I know enough myself! 

Well, the time was now, and as it turned out, it was fortuitous timing as my students were working on an assignment where ChatGPT would be useful. They were researching Hinduism, Christianity, and Islam in a jigsaw style assignment. The goal of this wasn’t necessarily to produce a written product, but to gain knowledge in one area, and then teach each other. I wanted to do this quickly – Religion for Dummies, if you will – so we could move on with our novel, Life of Pi, which speaks extensively about all three belief systems. As students started researching some of them quickly became overwhelmed by the sheer amount of information about these religions, and did not know where to start, or how to narrow down the breadth of knowledge available. We opened ChatGPT and simply asked it the questions I asked them: What are five significant rituals or traditions of Hinduism… what do Muslims believe is the reason for existence… what do Christians believe about the afterlife? Rather than sifting through pages and pages of documents trying to put together this information, students were quickly able to type in the exact question and generate a straightforward answer, even specifying that they would like the answer in “easy to understand language.”

Even as I type this, I recognize that this tool could easily become a crutch, an easy way out that replaces the act of research, the practice of critically evaluating and synthesizing information. These skills are essential, and it’s important that we teach them.  As I was showing the kids this tool, they were blown away by the ease of the task. One even asked, “Is this … cheating? It feels like I’m cheating.” I had to explain that the purpose of this assignment was not the act of research; rather, it was simply to quickly gather information and jigsaw shared data. In the article Rokhsareh and Hanieh shared, “The Future of Education: Artificial Intelligence based Remote Learning,” the authors list a number of ways AI can help students and teachers, and they list this kind of activity as one of them – the speedy finding and disseminating of data. But the whole conversation did bring up the tensions between using AI effectively and using AI as an easy way out. I haven’t yet reconciled these issues in my own mind, and thus, as I was sharing how to use the toll with the kids, I kind of felt like I was giving matches to an arsonist. Lol. 

Josh Brake, The Absent-Minded Professor, suggests three considerations to keep in mind as navigate this new tech. He bases his suggestions on other examples of new tech from the past, including the calculator, Wikipedia, and Google Search. 

1. “Make ladders, not crutches” – Here Brake discusses the same concern I felt when the kids were researching information today. He says, “aim to strike a balance that ensures that students know the fundamentals, but also train them to use the technology effectively to extend their capabilities.” The problem today was that I don’t know that all kids had actually mastered the fundamentals of research — thus the comment that this felt like cheating. Yet, it facilitated the lesson that I actually wanted to focus on in a much faster way. 
2. “Take AI output with a grain of salt” – Blake compares the tool to Wikipedia. You cannot automatically assume that the information is accurate. He shares an image I thought was interesting: 

Just like Wikipedia, we need to teach kids that they will need to verify information and confirm its accuracy. 

3. “Mind the algorithm and its pathologies.” I thought this was a particularly interesting piece of information as I had not thought about the algorithm. In my head I was kind of thinking of it as a neutral Google. But of course it’s not – Chat-GPT functions like any other algorithm. Blake cautions that we need to careful in our assumption of neutrality: “While we all have our own implicit biases, the tendency of the algorithm to surface these and combined with the overconfidence problem discussed above (doesn’t seem like the AI often expresses doubt) makes for a perfect storm that we should be wary of.” I need to spend more time playing around with the tech in order to talk to kids about this particular issue; it would be useful to generate examples of its bias. 

A final story to share: last week in class, after we had talked through Chat-GPT, I had my students run their inquiry essays through ZeroGPT and come show me the results. This is a 1500 word research paper; it is a grade 12 outcome. I explained to students that this will become standard practice for us as we move forward with AI technology and ensuring academic integrity. We discussed how ZeroGPT and other like sites are not always accurate, and how they can give false positives as well. (Sidenote: it is interesting  to read that ZeroGPT’s tagline is “Humans deserve to know the truth”!) It was interesting to see and hear their reactions. Some kids were immediately stressed about the false positives – the worriers who definitely did not cheat, but get stressed at the notion that someone might THINK they are cheating. Most were just interested to see what the site would generate. Finally, there were a couple who looked immediately nervous; one came to me right away and told me he had run his paper through a grammar checker (though he couldn’t remember the name of the extension), and he was worried it might come back with a high percentage. Sure enough, it came back at around 50% AI generated. He wondered if he might not take his essay home and rework those parts “fixed by the Chrome grammar extension.” I thought that might be a good idea, and he could maybe hand it in after the break after double checking his work. 

Advertisement