The companies are hiding behind this as a feature - that the algorithms are not supposed to be implemented without human review. I wonder how this "feature" will interact with implicit (and explicit) biases, or with the power dynamics between adjuncts, students, and departmental administration.

The companies are caught between a rock and a hard place in the decision whether students should be informed that their attempt was flagged for review, or not. We see that, if the student is informed, it causes stress and pain and damage to the teacher-student relationship. But if they're not informed, all these issues of bias and power become invisible.

I can't help but wonder if the emphasis on "content retention" and "compliance" that seems to be core to the authors' concept of learning doesn't make some bad assumptions: that learning is something that an instructor does to a student, and not something that students have agency over. This seems to me to be in extreme conflict with what might be even more inclusive practice: far less emphasis on the grade, more individual attention and greater emphasis on personal growth, less teacher control and more student agency. This is basic Freire stuff. Students aren't vessels to be filled.

It's worth noting that none of this is easy or efficient. If it was, we'd all do it. One must make a concerted effort to ensure students feel like they belong and that they're supported so that they can learn best. If that's not part of an instructor's job, then what exactly is the instructor's job?

minimize inequities = reduce harm; clear similarities to Hippocratic oath. See also Kaufman & Schipper (2018) Teaching With Compassion, p. xxiii

Emphasis on all. Frequently defensiveness kicks in when traditional teaching is called into question, because it does work, but just for some.

Although K-12 settings and children are referenced, I find the information in this document relates to all people -- students in all settings.