Mike, as you know this is a hobby horse of mine. IMHO you are asking all the right questions. One potential and partial solution to think about is what might be thought of as the "mobile device method". Let us assume that we can agree that students (or their parents for minors) own the data they create (this is not straightforward in itself as many vendors will claim they own the data that is analysed by their systems - there may be a way forward here in identifying "primary" and "secondary" data, "secondary" data being that produced after analysis and algorithmic processing). Systems can then give students visibility, or ideally choice, about what algorithms are applied to their data. A mobile device operating system asks you "app x wants to use your location. Is this OK?". We can do the same in education. We can alert students to a) which data is going to be processed (for example, the results of their end-of-unit tests, their age, their gender) and b) what results are going to be given to whom (recommendations for them to learn what they have not mastered, indications of individual and class-level weaknesses to teachers, aggregated test results to government). At the very least, this might encourage people to think about how their data is being processed and force some degree of transparency on all those analysing educational data.