One problem here is analogous to the separation of powers in government: education and testing are separate powers, and shouldn't be exercised by the same body.

One problem here is analogous to the separation of powers in government: education and testing are separate powers, and shouldn't be exercised by the same body. Educators ought to be doing everything they can to help students gain understanding, and such evaluation as is necessary for them to guide that process ought to be consequence-free, except to alert the educator to what next step is needed. Doing otherwise creates a needlessly antagonistic relationship between teacher and student, and causes all sorts of unhelpful incentives in both parties. Evaluation of the success of that education, on individual and institutional levels, ought to be done by people not involved in the process of education. Such testing should be infrequent enough that live, in-person proctoring would be no burden. There is no reason for high-stakes evaluation to be a constant feature of education. I don't think that'll happen anytime soon, but it seems like the right approach, to me.