A dataset on safety trust in Intelligent Domestic Robots collected from both humans and ChatGPT-4o.

The article consists of three studies, with data distributed across eight Excel files.All data were collected online through the Credamo platform.Study 1a_1: The raw data includes 1,293 participants, encompassing participant ID, demographic information, experience with smart devices, preferences for robots with different degrees of anthropomorphism, and responses to an initial 50-item scale.

Details of the scale items can be found in the article's appendix.Study 1a_2: The data storage format is consistent with Study 1a_1, except that it does not include preferences for different anthropomorphic robots or the initial 50-item scale. Instead, it contains detailed data from the traditional Automation Trust Scale (Aitems) and a self-developed 19-item Human-Machine Trust Scale (Bitems), with data from 433 participants.Study 1b: The raw data includes 139 participants, covering experimental groupings (group), demographic information, pre- and post-test willingness to use (Aitems), overall trust (Bitems), compliance trust (Citems), relational trust (Ditems), performance trust (Eitems), and human-machine trust (Fitems) scale scores.Study 2a: The raw data includes 720 participants, including anthropomorphism groupings, demographic information, compliance trust ratings across seven items under different height and camera visibility conditions, and risk perception ratings on two items.Study 2b: Data generated by GPT-4 API includes 12 conditions (matching the column names), with seven-item ratings collected under each condition and 30 data points collected in total.