Robert B. Cialdini

Are large language models (LLMs) susceptible to the same persuasive appeals as humans? We tested whether classic persuasion principles (authority, commitment, liking, reciprocity, scarcity, social proof, and unity) could induce three widely used LLMs (GPT-5 mini, Claude Haiku 4.5, and Gemini 3 Flash) to comply with requests to assist with the synthesis of regulated substances. Across 126,000 conversations, persuasion principles increased compliance from 35.3% (at baseline) to 51.3% (using any principle). Although LLMs are not human, these findings underscore their parahuman (i.e., humanlike) nature and reveal the risk of manipulation by malicious users seeking to circumvent safety guardrails.

Three mostly positive developments in academic psychology—the cognitive revolution, the virtual requirement for multiple study reports in our top journals, and the prioritization of mediational evidence in our data—have had the unintended effect of making field research on naturally occurring behavior less suited to publication in the leading outlets of the discipline. Two regrettable consequences have ensued. The first is a reduction in the willingness of researchers, especially those young investigators confronting hiring and promotion issues, to undertake such field work. The second is a reduction in the clarity with which nonacademic audiences (e.g., citizens and legislators) can see the relevance of academic psychology to their lives and self-interest, which has contributed to a concomitant reduction in the availability of federal funds for basic behavioral science. Suggestions are offered for countering this problem.

Böckenholt and van der Heijden’s results regarding compliance with insurance regulations—that the enforcement activities of a regulatory agency were relatively unpredictive of compliance—are consistent with findings from other domains (e.g., tax adherence), where personal factors and informal social controls have been shown to play a more significant role. However, the specific form of informal social control investigated in Böckenholt and van der Heijden’s study (the perceived approval/disapproval of friends and family) is not the only kind of informal social control that has proven effective in spurring compliance. Descriptive social norms, which involve perceptions not of what others approve but of what others actually do, also influence compliance decisions powerfully. Yet, the role of descriptive social norms in rule adherence is often underappreciated by governed and governors alike. The consequences of this relative lack of recognition are discussed within the arena of compliance with pro-environmental regulations and requests.