Answers are grounded in this paper and cite their sources.
Bots and large language model (LLM) agents can now complete online surveys convincingly enough to pass as real respondents, a risk long flagged in other fields but rarely addressed in public management research. Drawing on disheartening empirical findings from previous efforts to validate responses to online panel surveys, this research note offers a portrait of efforts to ensure the integrity of data collected by firms that survey citizens on behalf of Canadian federal departments and ministries, a case study likely representative of the challenges facing online survey practices across countries and levels of governments.
No takes yet. Share an insight, caveat, or question.
Morin et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: