How AI is rewriting the syllabus — practically

When five of my colleagues and I sat down at the start of last semester to plan how we'd integrate AI into our teaching, the conversation kept circling back to the same question: how do we use these tools without replacing what makes teaching valuable in the first place?

Six months later, we have an answer — and it's both more and less than we expected.

The breakthrough came when we stopped thinking about AI as a replacement for any part of teaching, and started treating it as infrastructure. Specifically: as a layer that handled the work we never had time for anyway.

For Dr. Menon, that meant generating differentiated practice problems for students working at different paces. For Prof. Sharma, it meant rapid-cycle feedback on draft essays — students got three rounds of feedback in the time they used to get one. For me, it meant being able to design a rubric for a project I'd never assigned before, and have a defensible version ready in fifteen minutes.

None of us are using AI to write our lectures. None of us are letting it grade final exams. But all of us are getting back somewhere between four and seven hours a week — time we're spending on the students who need us most.

The lesson, I think, is that the question isn't "should we use AI?" It's "what work do we do that an AI could do better, so we can do the work only we can do?" That second question is harder. But it's the one that actually changes how you teach.