Within the HRI community, there is an increasing emphasis on adopting a paradigm shift that focuses on real-world contexts—where uncertainty and unpredictability challenge traditional research paradigms. This enables robots to be deployed in real-world applications to support humans, where these robots either directly interact with users—such as serving as receptionists, tour guides, home assistants, or companions—or execute tasks under an operator's remote supervision and control, such as in space or underwater explorations. As robots face these uncertainties, complexities, and dynamics in such challenging deployment contexts, failures can arise from human, robot, task, environment, interaction, or scaling aspects.
This workshop continues the trajectory of PubRobFails and ExactingHRI, uniting their communities under a shared focus on real-world deployment challenges and learning from failures across diverse contexts. We invite researchers, designers, and practitioners to reflect critically on the technical, social, and organizational challenges of deploying robots in public, private, and extreme environments. We encourage open discussion of both successes and failures, including stories that rarely make it into traditional conference sessions—those involving unexpected breakdowns, messy real-world data, and valuable lessons learned.
This workshop offers a platform to explore what happens when robots meet reality, and how the HRI community can grow by sharing not only what works, but also what does not. Embracing a culture of "failing forward", we aim to transform setbacks into stepping stones for progress, fostering a culture of error acknowledgement, transparency, and actionable insights that can unlock the full potential of HRI research.
As discussions in the PubRobFails and the Exacting HRI workshops revealed, robotic systems face both predictable failures (e.g., battery depletion, network loss) and unpredictable ones (e.g., crowds swarming a robot, operator overload, stakeholder conflicts). Minor oversights in planning or communication can propagate into major execution breakdowns, often driven more by human factors—participants, bystanders, or team members—than by the robot itself. Participants also emphasized the importance of redundancy, safety buffers, and recovery time, as well as recognizing that failure analysis extends beyond operation into data interpretation and reporting. These failures occur across domains, whether a service robot in a shopping mall, a teleoperated robot in a nuclear plant, or an assistive robot in healthcare. Building on these insights, this workshop will explore:
Technical, human, organizational, and environmental factors that cause deployment challenges.
Strategies for anticipating and mitigating failures.
Methods for analyzing failures and turning them into actionable design insights.
Cross-domain lessons: What can different environments teach social and field robotics?
Ethical, safety, privacy, and governance considerations.
Topics of interest include (but are not limited to):
Human unpredictability, stakeholder alignment, and cultural/social factors.
Robust perception and planning under uncertainty.
Error detection, recovery, prevention, and trust repair.
Long-term, in-situ evaluation and data quality challenges.
Governance, safety, privacy and ethical frameworks.