Helen Hastie on the future of human-robot relations

Episode Summary

Executive Summary: Helen Hastie, professor of human-robot interaction and head of Edinburgh’s School of Informatics, explains how useful robots must be task-specific, trustworthy, transparent, and designed with human behavior in mind. The conversation traces her career from early dialogue systems to social robots used in coffee service, education, healthcare, and hazardous work, while emphasizing responsible innovation, regulation, and realistic expectations.

Main Topics: Human-robot interaction and trust (Priority: 5/5): Hastie argues that robots should be designed for appropriate trust: not over-trusted, but also not under-trusted when they outperform humans in specific tasks. The limits of sci-fi robot expectations (Priority: 5/5): She challenges the media image of all-purpose humanoid robots, stressing that today’s robots are narrow, task-based systems rather than general intelligence machines. Evolution from dialogue systems to generative AI (Priority: 4/5): The interview follows her early work on spoken dialogue systems and IVRs, showing how these systems evolved into today’s data-heavy, end-to-end AI assistants. Socially aware robots in real-world settings (Priority: 5/5): Projects like the Robo Barista, empathic tutor, and triage nurse illustrate how robots can use conversation, emotion cues, and explanations to improve adoption and usability. Testing, deployment, and public acceptance (Priority: 4/5): The National Robotarium and public engagement work show the importance of real-world testing environments and measuring public attitudes over time. Responsible AI, regulation, and future risks (Priority: 5/5): Hastie emphasizes guardrails, transparency, auditability, and protection against hacking or misuse, warning that innovation should not outpace safety thinking. Robots for independent living and social good (Priority: 4/5): She envisions robots helping older people live independently, doing dirty, dangerous, or repetitive tasks, and becoming a normal part of daily life.

Key Arguments: Robots are currently highly capable only in narrow, specialized tasks; general-purpose sci-fi robots are still far away. Robot appearance should fit the job: human-like forms can help in homes, but functional designs are better in industrial settings. Appropriate trust is essential: users should trust highly accurate robots when warranted, but remain vigilant in safety-critical contexts. Transparency improves acceptance, especially when robots explain decisions, such as in medical triage. Emotional and conversational cues can improve interaction, as seen in the coffee robot, tutor, and apology studies. Real-world deployment matters because controlled lab success does not guarantee public adoption or long-term use. Responsible research and regulation should be embedded from the beginning of design, not added later. Robots are most valuable when they help people in dirty, dangerous, or socially beneficial work and support independent living.

Data Points: Robo Barista trial duration: 6 weeks - Public interaction study measured attitudes over time How May I Help You? timing: 2001–2002 - Early interactive voice response project at AT&T Research Labs Year of Royal Society Summer Exhibition: 2019 - Public engagement exhibit on robots for dirty and dangerous jobs Year became academic co-lead of National Robotarium: 2020 - Joint initiative between Heriot-Watt and University of Edinburgh Year became head of School of Informatics: 2023 - Return to the University of Edinburgh Family age mentioned for father: 95 - Example of older adults adopting technology Family age mentioned for mother: 90s - Example of older adults and robot-assisted independent living

Pivotal Quotes: "We are far from that. We do have some robots that can perform very, very well. On specific tasks, but these tend to be very narrow tasks." — Helen Hastie: On the misconception that robots are already general-purpose, human-like assistants "It's good that robots look a bit like us, but not too much like us." — Helen Hastie: Explaining the uncanny valley and robot design "What is important, I believe, is that we build in these systems a level of transparency and an audit trail so that we can understand the decisions that are being made and the consequences." — Helen Hastie: On regulation, accountability, and safe deployment

Implications: Listeners should expect robots to become more common, but mainly as task-specific, transparent tools that augment humans. For industry, the message is to prioritize usability, trust calibration, and safety-by-design over flashy humanoid form factors.