human-AI interaction ethical AI prosocial systems empathy
This research examines whether AI can provide empathetic emotional support without being sycophantic.
While preferred by users, overly agreeable LLMs have been shown to reduce users’ critical thinking and sense of accountability. Jessica will be testing whether introducing “prosocial friction” into an AI’s empathic responses can preserve emotional validation while promoting more thoughtful engagement.
→ This research will be conducted in collaboration with The Pennsylvania State University’s Rock Ethics Institute.
→ Data collection for the initial study is expected to begin Fall 2026, pending IRB approval.