I have built check-ins that learners were glad to get, and I have seen dashboards that only made people feel watched. The difference was never the technology. One said I noticed you, how is it going. The other said I am counting your clicks.
Checking in is not checking on. Support that respects the learner assumes competence and offers help; surveillance assumes the worst and measures compliance. The design choice between them shapes whether a learner returns.
Autonomy-supportive structure
Structure and autonomy are not opposites. Adults sustain effort when the structure supports their agency rather than replacing it — when they have real choices inside a clear frame, and someone who notices when they act (Ryan & Deci, 2000). Accountability designed this way feels like care. Designed as monitoring, it feels like distrust, and distrust is a reliable way to lose an adult learner.
Checking in is not checking on. Accountability that respects the learner assumes competence and offers help.
The restart: the gap is data, not debt
Adults leave and come back. Life intervenes, and a gap opens in the course. The punitive design treats the gap as a debt to be paid down with shame. The humane design treats it as data: something changed, here is the on-ramp back. The learners who returned after a long absence taught me to build for the return rather than to penalise it — because for most adult learners the return, not the uninterrupted run, is the normal case.
Why AI raises the stakes
AI makes fine-grained surveillance cheap: every keystroke loggable, every pause flagged. That temptation should be resisted precisely because it is easy. The same capability can instead power gentle, timely nudges and easy restarts. Cheap monitoring is not the same as good support.
The move
Design one visible, low-stakes restart path into a course — a clear way back after a gap that carries no penalty. Build for the return, because it is coming.