This is an assessment strategy that is being used across Computing to mitigate the use of AI. It is being used in large class sizes to maintain academic integrity.
We (in Computing) have mostly come to the realization that take-home assignments are no longer a viable means to assess with the level of technology that currently exists. What many of us have resorted to is (1) keeping the traditional take-home assignments; (2) dropping their weight in the course drastically (sometimes entirely); and (3) assessing their understanding of a strong submission through in-person quizzing. It takes away from being able to ask about other things during in-person assessments, but seems to be the only way to maintain integrity with large class sizes, given the impact GenAI is having.
It mitigates by necessitating understanding of the material without the aid of AI during the primary assessment (in-person quiz). Studying an AI generated assignment does not offer near the same level of understanding (and thus performance on the quiz), as doing the assignment authentically.
The assignments are large in scope and directly speak to the CLOs. Removing them entirely from the course would make hitting those CLOs much harder.
The assignments could be almost entirely auto-generated with minimal effort.
Yep. I'm doing this for the 3rd year in a row now. If I had a magic wand, I'd have a unique quiz for each student that included their own code on the quiz. Logistically I haven't been able to imagine how this could work (yet).
Additional Information
N/A. It's more of a general assessment strategy.
Some have the assignments worth 0%, but I find this leads to more students completely skipping that portion.
Classes where large assignments are historically effective, but now subject to GenAI completion.
All of them (it's a generic strategy).