This episode features Chris Chapman (Executive Director, Quant UX Association) and his take on LLM-generated “synthetic survey responses.” He lays out two paths to learn what people think: ask real people directly, or build a pipeline that generates “responses” using models, prompts, and “digital twins.”

Chris explains why he believes synthetic responses are never something you can certify in advance, why ideas like random sampling don’t map neatly to LLM outputs, and why surveys are better understood as motivated communication shaped by wording and context.

Watch the full webinar and download Chris’s slides here.