Teachers everywhere are currently realizing that students now have an incredibly new, powerful tool for creating unique essays within minutes that are decent quality, and, most notably, not their own original work. If you haven’t already, visit this new chat bot and go take a look. Ask it a few questions. Plug in your writing prompts. It’s really something. Currently, this is free and open to everyone. It’s highly unlikely that you have many students who aren’t aware of this new technology, so you should familiarize yourself with it, too. So, for those of us that have heavy writing elements in our courses, now is the time for us to take a close look at how we are assessing student progress in our classes.
When I enter one of the AP® World learning objectives, it produces a three-paragraph response in under a minute. I asked it to “Explain the causes and effects of the ideological struggle of the Cold War,” and here’s what it gave me:
The trouble with this is that our normal go-to plagiarism detectors won’t always catch this as it’s a unique response. The chat bot isn’t pulling this from somewhere on the internet- it’s creating it on the spot. There is an online detector that generates a probability that a written piece is fake or authentic. Find it here. When I put the above response to the Cold War prompt into this detector, it tells me it is 99.98% fake.
How do we combat AI-generated plagiarism?
There are few tricks that can be helpful in determining if a written assignment was generated using AI:
- Use the GPT-2 Output Detector mentioned above
- Check revision history on Google Doc essays- zero revisions on a written assignment is a definite flag
- Ask students to verbally summarize or present their work
- Ask questions about specifics in each essay/paper
The inherent problem with trying to come up with a way to figure out if an assignment was plagiarized is that it’s a reaction to a problem. We want to prevent this kind of thing from happening rather than spend valuable time trying to catch cheating. Once a student has submitted work that wasn’t their own, we’ve already lost that student in the learning process.
Creating Written Assignments that Discourage the Use of AI
Full disclosure: I am not in the camp of “the essay is dead” in terms of secondary education. I believe the writing process is a crucial part of learning to think clearly and critically about the world. While it may initially seem like this new chat bot has ruined much of our ability to give written assignments, that is actually far from reality. It’s more important than ever that we teach, and assess, the actual process of writing, and that we do it right in our classrooms. Part of any longer written assignment should include the pieces of the entire process: outlines, drafts, annotated bibliographies, etc. Have students practice writing thesis statements. Practice identifying the line of reasoning within another written piece. These kinds of things are just as important as a finished essay for high school students.
Of course we also want to collect short, written responses from students, too. This is where in-class, hand-written responses may need to make an occasional come back if you’ve switch everything online.
Lean In to AI with a “Reverse DBQ”
This specific example is for history classes, but the overall idea can be expanded to other subject areas, too. With any new technology, resistance is (usually) futile. So, instead of trying to resist ChatGPT and other AI language bots, it’s time to figure out how to lean in and incorporate them into our classrooms. Here’s one of my ideas for my World History students:
In small groups, give students a prompt like the one used above (Explain the causes and effects of the ideological struggle of the Cold War). You can use different prompts for different groups or the same for each depending on your resources. Have students use ChatGPT to generate a response to the prompt. Then, give each group a set of primary and secondary sources. These should be a mix of relevant and irrelevant sources. Each group needs to read or examine each source and identify which ones could be used to support certain points in the chat bot’s essay. Students should first identify the relevant sources, then have them write a short explanation for each relevant source about its connection to the AI-generated essay. What points does it support or refute? Is there any important information in any of the sources that’s missing from the AI-generated essay? How could the essay be expanded or polished up?
This activity has students working with written material, primary sources, and the dreaded essay-writing bot in a way that encourages thinking rather than regurgitation. Getting students to become critical of the chat bot is my goal, because that requires them to engage with the content of the class in a meaningful way. And hey, as teachers, we also need to figure out how these new chat bots can make our lives easier, too.
Take a closer look at a “Reverse DBQ” here.