What To Do If You Suspect AI Misuse

I've talked a lot about how students and teachers can use AI to improve their productivity, build upon ideas, or get feedback. But there are also plenty of ways that AI can be misused. When ChatGPT was first released, some students began using it to write essays for them. It led to a teacher panic and a desire to block the website and any other future AI tools. But blocking the tools can only go so far. Students can still access AI from home or from their phones. So what should we do as teachers?

Well...as with anything, it is all about education. We need to teach students about how to use AI ethically. We need to discuss the line between support tool and academic dishonesty. Ideally, we would have done that before the students ever got access to tools like ChatGPT, but unfortunately, that didn't happen. So now we are trying to backtrack and teach these important skills after the students have already been rewarded with good grades for having AI write for them. It's like trying to train your dog to stay off the counter after he got a whole steak.

I absolutely advocate for focusing on educating our students about how to use AI responsibly, but I am also realistic and understand that we need to find ways to provide consequences for students who misuse the technology. So the question that keeps coming to me from teachers is...how do I know if they used AI to write this?

While undeniable proof that a student used AI is rare, here are some techniques I use to assess whether it is "more likely than not" that the work isn't the student's own:

Use Your Gut

Ask yourself the following questions:

• Does the writing style differ significantly from previous work by the same student?

• Are there any words that seem advanced or unusual?

• Does the writing answer the question posed with specific details and/or citations?

• Are there unusual or advanced concepts introduced without explanation?

• Is there a lack of personal experience or unique perspective in the writing?

Use a Technology Tool - Brisk

I like to use Brisk Teaching's "Inspect Writing" feature to quickly see a replay of how the document was created (note that the AI detector mentioned in this article is now only available in the paid Pro version of Brisk). It shows me how long it took the student to type the document and whether there were any large copy/paste elements. Here are a few examples:

No AI usage:

Notice the time it takes to write the content and the fact that there are changes made as they write. This tends to indicate that a real person is writing.

Typing what AI wrote:

Notice how quickly this content is typed and that nothing is changed from start to finish. This tends to indicate that a real person is typing, but that they are copying from something else (possibly AI).

Copying and pasting entirely from AI:

Notice how there is one large copy and paste element. This indicates that a person did not type the writing and that it could be copied from AI.

Copying pieces from AI and adjusting them:

Notice how there are copied and pasted elements that then have a few words changed. This tends to indicate that the copied content was written by someone else (possibly AI) and then a person made adjustments to make it sound more like them.

Talk to the Student

If you suspect misuse...have a conversation with the student. Consider asking some of the following questions:

• Did you use AI in any way to complete this assignment?

• Can you summarize your writing in your own words?

• Why did you choose this topic/argument/perspective?

• Can you explain how you organized your ideas and structured your writing?

• Can you explain why there are large elements that have been copied and pasted?

• Can you explain what [specific word/sentence/concept] means?

• Can you explain where you found your sources?

• Can you explain why you did not include citations?

If a student admits AI misuse - of course there can be consequences, but make sure that they understand why their actions were wrong and provide examples of how they could use AI in a more ethical way in the future (perhaps to get feedback on something they wrote or to help with ideas if they get stuck). Remember that this technology is new and our students are young.