Campbell, Sinéad. “‘Our Assumptions Are Broken’: How Fraudulent Church Data Revealed AI’s Threat to Polling.” The Guardian, Guardian News and Media, 28 Mar. 2026, www.theguardian.com/technology/2026/mar/28/how-fraudulent-church-data-revealed-ais-threat-to-polling.
Summary: Sinead reminds readers of a previous controversy where survey polling data collected by the Christian Church in Britain that was used to generate popular news stories that Gen z was returning to church and religious practices more so than previous generations was re-evaluated and turned out to be wildly inaccurate. This is because many online accounts that were surveyed by the Church turned out to be ran by AI programs created to generate a more positive data set than what was predicted.
Why this matters: With the wide prevalence of AI, and it’s deep integration into online internet systems, and especially with how people can generate social media accounts using rudimentary AI LLM programing, it has become much harder to conduct research studies that utilize demographic polling. Internet-gathered data, especially if done anonymously has become much harder to prove if the collected data matches reality.
Leave a Reply