Several papers follow a common progression: first predicting human personality traits from text, then testing whether an LLM can reliably simulate a specified personality profile.
This research connects to conversational AI personalization, controllable LLM personas, and human-centered AI.