As an online graduate instructor, I can usually tell when my students are using ChatGPT to complete their assignments. While we allow AI use in our program, we simply ask students to acknowledge it, but many don’t. But there are some ways to know that students used AI-generated content in their work. When I run my own assignments through ChatGPT, I immediately notice the telltale signs: neatly bolded headers (my first clue!), perfectly formatted bullet points (often ending with periods), flawless grammar, and impeccable spelling. Everything looks extremely polished, usually TOO polished. This formatting creates a great user experience, but it lacks the natural imperfections of human work.
What I find particularly revealing is how AI handles our reflection components, which we specifically designed to verify student learning and help us figure out if students are over relying on AI. Instead of informal, personalized thoughts, AI-generated reflections tend to be generic and perfectly structured, as well as filled with those unmistakable em dashes that AI loves so much. There’s no personality, no authentic voice when students write things like “Help me become more aware of how I organize complex information” (who actually talks like that?). Rather than relying on ineffective AI detectors or immediately jumping to academic integrity violations, I recommend having open conversations with students about their process. Don’t ban AI outright! They’ll use it anyway and just hide it from you. Instead, guide appropriate usage through conversations with your students.
Watch my video for more: