The trend of bringing large language models to empathy and empathic relatedness is a game changer. The question is not whether the generative AI can be empathic, but the extent to which the designers are able to distinguish responsiveness from “stroking one’s ego,” sycophancy (servile flattery), and the extent to which prospective clients decide to engage (both open questions at this date (Q1 2026)).
Caruth (1996) concisely defines trauma in terms of an experience that is registered but not experienced, a truth or reality that is not available to the survivor as a standard experience. The person (for example) was factually, objectively present when the head on collision occurred, but, even if the person has memories, and would acknowledge the event, paradoxically, the person does not presently experience it as something the person experienced in a way that a person standardly experienced the past event. The survivor experiences dissociated, repetitive nightmares, flashbacks, and depersonalization.
Empathy Lessons 