Two clocks are ticking: the model clock (measured in months, sometimes weeks) where AI capabilities compound, and the institution clock (measured in semesters and committees) where your university grinds forward.
Thanks! Likely shape: clear notices wherever AI is used; risk/impact assessments; consent/opt-out for sensitive analytics; strong privacy/security and data minimisation; human-in-the-loop with a right to review/appeal; provenance checks for assessment/research; audit logs/incident reporting; and portability/exit clauses for vendors. Translation: show your working, protect students, keep humans accountable, and design for exit.
“AI-in-education rules” - what do you reckon those will look like? Fascinating piece by the way. Definitely can see a lot of this already beginning.
Thanks! Likely shape: clear notices wherever AI is used; risk/impact assessments; consent/opt-out for sensitive analytics; strong privacy/security and data minimisation; human-in-the-loop with a right to review/appeal; provenance checks for assessment/research; audit logs/incident reporting; and portability/exit clauses for vendors. Translation: show your working, protect students, keep humans accountable, and design for exit.