Any norm like this should be based on data. There is evidence that use of AI for a task neg correlates with how much they understand it. This norm implies AI makes tasks more efficient, but again, data suggests it make one less efficient but give the impression otherwise.

If one is opposed, based on the wide range of moral arguments for integrating AI into everything, why would one "try it out"? The any apparent utility is almost irrelevant. I think far more pressing is that boosters face those args to be "informed".

As the greatest Canadian once put it.