Comment by ben_w

We could call this "reinforcement learning from human feedback" (RLHF) :)

https://en.wikipedia.org/wiki/Reinforcement_learning_from_hu...