As someone who works on an adaptive tutoring product, I am challenged and inspired by some of the questions raised here.

I believe algorithmic education has its place so long as it does not overreach. There are clear benefits to adaptive tutoring in embedding core skills and concepts and supporting students to develop at their own individual pace. For instance, I think efficiency is a crucial goal in the teaching of mathematics because of how woefully inefficient existing models of curriculum and assessment are. Adaptive tutoring can get students up to speed on foundational material with a level of precision that even the best-willed teacher can not. An active goal of tutoring services should be to enrich the student-teacher relationship and teachers’ time is better spent on deeper engagement with students.

Real-time reporting is the sister of algorithmic sequencers, and your analysis applies equally well to them (if not more so, since it is the reports that we push onto parents and educators). It is critical that progress data is not judge, jury and executioner when it comes to evaluating students’ effort and progress. Rather, it should give space for context that only parents and teachers can provide. I’m optimistic that technology will evolve to take into account not only students’ raw performance metrics, but also their emotions, their preferences and their broader circumstances. Even then the experience must allow for contextual interpretation of data and an empowerment of teachers.

Much depends on the intention of the underlying service provider: technology must lie in servitude of our pedagogical goals and unfortunately many providers miss this point entirely.