Back to today’s briefing

Story perspectives

Study Reveals AI Threatens Integrity of Polling Data

11/18/2025

47 3

Summary

Story summary

A Dartmouth University study shows large language models pose a threat to online survey research.
Led by Sean Westwood, the researchers show an AI tool can mimic human responses and evade detection at 99.8%.
The study warns AI-generated responses could corrupt public opinion polls and influence election outcomes.
Westwood calls on scholars to innovate data collection methods to counter AI manipulation and protect democracy.

1 of 1

Story summary

A Dartmouth University study shows large language models pose a threat to online survey research.
Led by Sean Westwood, the researchers show an AI tool can mimic human responses and evade detection at 99.8%.
The study warns AI-generated responses could corrupt public opinion polls and influence election outcomes.
Westwood calls on scholars to innovate data collection methods to counter AI manipulation and protect democracy.