The case for calling AI model ‘personas’ ‘egos’ instead

A user proposes replacing ‘persona’ with ‘ego’ when discussing AI models, arguing that ‘persona’ suggests a superficial mask over a true self. They suggest prompts can activate different latent egos in base models, while assistant models often present one active ego. Another user says they’ve used ‘ego’ to describe part of what forms in a posttrained model, but that the term doesn’t capture all of it.