Proceedings of the Annual Meeting of the Cognitive Science Society

Despite substantial charitable giving, donations often fail to maximize impact. While a variety of persuasive strategies can increase donations to effective charities, their success depends on individual differences. Large Language Models (LLMs) offer a powerful solution to this problem by dynamically personalizing persuasive strategies. In a pre-registered experiment (N=1952), we tested whether personalized LLM conversations could increase donations to the Against Malaria Foundation (AMF), rated one of the world's most effective charities. Participants allocated $1 between their favorite charity and AMF after being assigned to either: (1) a personalized persuasive LLM conversation, (2) a static LLM-generated persuasive message, and (3) a control conversation. Personalized LLM conversations significantly increased donations to AMF by 46.6%, outperforming the static message (28.7% increase). Personalized LLMs also shifted moral attitudes about charitable giving. Our findings highlight the potential of AI-driven personalization to enhance effective giving and provide new insights into the psychology of charitable persuasion.