Poll Variables Reveal AI Polling Limits

Publishing the full profile variable list (demographics, political behavior, religion, income, etc.) lets researchers and the public check whether synthetic respondents mirror real‑world diversity and response patterns. Making these variable lists standard practice would expose where AI‑generated samples succeed or fail and help regulators and newsrooms decide how much to trust synthetic polling. — Transparent publication of survey variable schemas is a simple, high‑leverage policy and scholarly norm that clarifies the limits of AI in reproducing public opinion and protects the integrity of polling.