Every teacher has read that paper — the one that sounds a little too polished, a little too smooth for the student who wrote it. That instinct isn’t paranoia. It’s a signal that the assignment itself has stopped measuring what it was built to measure. AI writing tools didn’t break your grading system by accident. They exposed a weakness that was always there: too many assignments reward a finished product, not the thinking behind it.
This course doesn’t hand you another AI detection tool or a ban policy that students route around in a week. It teaches you to redesign the assignment from the ground up, using a clear framework for deciding when to design AI out of a task and when to build it in on purpose. You’ll rebuild written assignments around real checkpoints, design tasks where AI collaboration is documented and visible, and even construct a subject-specific AI assistant your students can learn from directly.
By the end, you won’t be guessing whether a piece of work is “real.” You’ll have a repeatable system for building assignments that hold up no matter how good next year’s AI tools get — because they’re built around human judgment, not just human output.
