Dataset on Public Preference for Utilitarian Decision-Making by AI in Moral Dilemmas

Study 1. Participants were recruited in 2024 via the Credamo platform (https://www.credamo.com/), yielding a total of 478 responses. After excluding four participants who failed the attention check, the final valid sample comprised 474 individuals (309 female), aged between 18 and 45 years (M = 29.65, SD = 5.75).Study 2. Participants were similarly recruited in 2024 via Credamo, resulting in 290 valid responses. Participants were randomly assigned to either an anxiety induction group (n = 143; 71 female; age range = 19–42 years; M = 28.97, SD = 5.17) or a control group (n = 147; 68 female; age range = 18–45 years; M = 29.18, SD = 5.45). There were no significant differences between groups in terms of gender distribution (p = .74) or age (p = .73).Data analysis for both studies was conducted using SPSS 27 on a MacBook Pro. The datasets were structured in wide format for repeated-measures designs. The primary analysis involved repeated-measures ANOVAs. In Study 1, correlation analyses were also conducted. In Study 2, paired-sample t-tests were used to evaluate anxiety-related materials, and independent-sample t-tests were employed to verify the effectiveness of the anxiety induction manipulation.In the dataset of Experiment 1, Human-Utilitarian, Human-Deontological, Robot-Utilitarian, and Robot-Deontological conditions were used for repeated-measures ANOVA, with the Oxford Utilitarianism Score, robot anxiety, robot acceptance, and robot anthropomorphism being scale scores, along with the computed "Utilitarian Human-mAI" and "Deontological Human-mAI" representing the rating differences between humans and robots under utilitarian and deontological decision-making conditions for assessing interaction effect sizes.In the dataset of Experiment 2, Human-Utilitarian, Human-Deontological, Robot-Utilitarian, and Robot-Deontological conditions were used for repeated-measures ANOVA, with Group referring to the anxiety induction group and the control group (non-anxiety induction group), while anxiety level and anxiety cause were self-reported by participants after reading anxiety/non-anxiety induction materials.Anxiety Material Evaluation Dataset, where A represents anxiety material items and U represents non-anxiety material items. The anxiety material score is calculated as the mean of anxiety material items, while the non-anxiety material score is calculated as the mean of non-anxiety material items.