Teaching used to be one of those rare jobs that felt genuinely rewarding, even with terrible pay and zero job security. An Earth science instructor who's been teaching part-time for years says working with students was addictive in the best way. That was before ChatGPT.
Now, especially in asynchronous online courses, the experience has become mostly miserable. These classes were already harder than in-person teaching because you can't see when students are confused or keep them accountable in real time. Without scheduled class meetings or face-to-face interaction, students have always been more likely to disengage and fall behind.
Generative AI has amplified that problem dramatically. When students can feed assignments directly into ChatGPT, the already fragile connection between instructor and learner gets even weaker. The instructor can't tell who's actually learning and who's just submitting AI-generated work.
This matters because it's a preview of what's coming for anyone creating educational content or training materials. The traditional signals that someone is learning, understanding, struggling, are all breaking down. If you're building AI tools for education or corporate training, you're going to need new ways to measure actual comprehension versus AI-assisted completion.
The bigger question is whether we're optimizing for the wrong thing entirely. Maybe the future isn't about preventing AI use in learning, but completely rethinking what education looks like when everyone has an AI assistant. That's a much harder problem to solve than just detecting ChatGPT output.
