> Imagine randomized control trials where we measure how watching a specific movie impacts its viewer’s empathy, altruism, or motivation.

The problem is that as soon as you choose those categories, you encode the assumption that "elevating" art should impact the viewer in those specific ways (and, almost certainly, you have in mind a direction in which they should be impacted).

If we say that we want to promote specific pieces of art because we expect them to instill the right values in the general population, well, history has a lot of lessons to teach us about how that could end up.