AI爱？人工智能生成的配图对慈善捐赠的影响。AI and Love? The Impact of AI-Generated Illustrations on Charitable Donations.

本研究旨在系统探究人工智能（AI）生成图片对慈善捐赠的心理影响机制及其边界条件。随着生成式AI技术在公益领域的快速应用，其效率优势与真实性、伦理风险并存。当前研究主要证实AI生成内容可能削弱捐赠意愿，但其心理路径及关键调节因素（如受助情境紧急性与机构使用动机）尚未得到系统整合检验。 基于此，本研究旨在系统探究人工智能（AI）生成图片对慈善捐赠的心理影响机制及其边界条件。研究通过Credamo平台招募531名中国成年参与者，采用被试间实验设计。研究1检验图片类型（真实图片/AI生成图片）对捐赠意愿和捐赠金额的直接影响；研究2纳入情境紧急性调节变量；研究3a与3b在研究1、2的基础上，分别在非紧急与紧急情境下，检验慈善机构动机类型的调节作用和信任的中介作用。研究采用标准化的图片与文字材料操纵自变量，并使用已验证量表测量信任、捐赠意愿与金额等变量，综合运用方差分析、非参数检验及基于Bootstrap的有调节中介模型进行数据分析（PROCESS）。本研究预期构建一个有调节的中介模型，系统揭示AI生成配图对慈善捐赠的影响机制及其边界条件。具体而言：在非紧急情境下，AI生成配图（vs. 真实图片）预期通过削弱对慈善项目的信任而降低捐赠意愿与金额，且该中介路径受动机类型调节——道德动机声明（如保护隐私）可缓冲信任受损，而工具动机声明（如降低成本）则会强化负向影响；在紧急情境下，AI生成配图的直接负向效应预期消失，且调节模式发生反转：工具动机下的AI使用可能通过增强对项目执行效率的信任而促进捐赠，道德动机下则无显著差异。该结果将阐明AI生成内容在慈善传播中“何时（情境紧急性）、为何（信任中介）、对谁（动机类型）产生何种影响”的完整作用路径。 This study aims to systematically explore the psychological impact mechanism and boundary conditions of artificial intelligence (AI)-generated pictures on charitable donations. With the rapid application of generative AI technology in the public welfare field, its efficiency advantages coexist with authenticity and ethical risks. Current research has mainly confirmed that AI-generated content may weaken donation intention, but its psychological path and key moderating factors (such as the urgency of the assistance situation and the usage motivation of the organization) have not been systematically integrated and tested. Based on this, this study aims to systematically explore the psychological impact mechanism and boundary conditions of AI-generated pictures on charitable donations. The study recruited 531 adult Chinese participants through the Credamo platform and adopted a between-subjects experimental design. Study 1 tested the direct impact of picture type (real pictures/AI-generated pictures) on donation intention and donation amount; Study 2 incorporated the moderating variable of situation urgency; on the basis of Studies 1 and 2, Studies 3a and 3b tested the moderating effect of the motivation type of the charity organization and the mediating effect of trust in non - urgent and urgent situations respectively. The study manipulated the independent variables using standardized picture and text materials and measured variables such as trust, donation intention, and amount using validated scales. Data analysis was conducted using a combination of analysis of variance, non - parametric tests, and a moderated mediation model based on Bootstrap (PROCESS). This study expects to construct a moderated mediation model to systematically reveal the impact mechanism and boundary conditions of AI-generated pictures on charitable donations. Specifically, in non - urgent situations, AI-generated pictures (vs. real pictures) are expected to reduce donation intention and amount by weakening trust in the charity project, and this mediating path is moderated by the motivation type - a moral motivation statement (such as protecting privacy) can buffer the loss of trust, while an instrumental motivation statement (such as reducing costs) will strengthen the negative impact; in urgent situations, the direct negative effect of AI-generated pictures is expected to disappear, and the moderating pattern is reversed: the use of AI under instrumental motivation may promote donations by enhancing trust in the project's execution efficiency, while there is no significant difference under moral motivation. The results will clarify the complete action path of "when, why , and for whom" AI-generated content has what kind of impact in charitable communication.