Reproduce “Value Alignment and Trust in Human-Robot Interaction: Insights from Simulation and User Study” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Shreyas Bhat, Joseph B. Lyons, Cong Shi, X. Jessie Yang

The research investigates how value alignment between humans and robots influences human trust in human-robot interactions, demonstrating that task risk significantly moderates this relationship. It shows that robots capable of adaptively learning human preferences consistently foster higher trust and lower perceived human workload compared to non-adaptive approaches.

Key insights

  • The relationship between value alignment and human trust in robots is moderated by task risk; misalignment significantly reduces trust in high-risk scenarios but has less impact in low-risk ones.
  • Robots capable of adaptively learning human preferences through Inverse Reinforcement Learning (IRL) foster higher levels of human trust compared to non-adaptive or fixed-objective robots.
  • Specific combinations of human and robot value priorities (e.g., health vs. time) lead to distinct patterns of trust evolution across varying risk levels, identifying four behavioral regions.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.