This OSF project contains the data and analysis files for the study described in the chapter preprint: Empathic AI Will Undermine Human Kindness. This chapter is to appear in A. Perry and C. D. Cameron (Eds.) Empathy and Artificial Intelligence: Challenges, Advances, and Ethical Considerations. This study used behavioral measures to explore the impact of empathic AI on offloading empathy. Participants imagined an acquaintance experiencing a tough time and were told (1) nothing further, (2) that the acquaintance had already turned to ChatGPT for emotional support, or (3) that the acquaintance had already turned to another person for emotional support. To measure empathic behavior, participants wrote questions to the acquaintance about their situation and completed a perspective-taking task (Reading the Mind in the Eyes Task; RMET). Participants who learned that the acquaintance had gone to ChatGPT beforehand not only reported more desires to offload to LLMs, but also wrote fewer empathic questions and scored lowest on the RMET.