AI being used to participate in pay-for opinion polling at scale? Shocked I tell you! Shocked! https://www.theguardian.com/technology/2026/mar/28/how-fraudulent-church-data-revealed-ais-threat-to-polling
Oh, yeh, LLMs generate outputs designed to be plausible extensions of the input... so if the input is a survey and there's *any* implicit push in the survey, even unintentional, that's what the LLM is going to generate.


