Privacy – Consent

Consent in human-robot interaction sits at the intersection of privacy, safety, and social acceptability. It concerns who agrees to what kind of data collection or interaction, under which conditions, and for whose benefit. In practice, consent is rarely a single clear event. It is often distributed across time, context, and people: users interacting directly with the robot, bystanders who are affected without engaging, and institutions that deploy the system. This makes consent a persistent design and ethical challenge rather than a one-time checkbox. Within the broader value of privacy, consent is closely tied to questions of awareness and control over data flows. Robots operating in everyday environments may continuously record, infer, or transmit information, which raises the issue of whether everyone affected has meaningfully agreed to this. The setting strongly shapes the perception of consent: a robot used at home is experienced differently from one deployed in public or semi-public spaces like hospitals or care facilities. In these environments, bystanders may be monitored without having explicitly opted in, even though their presence is unavoidable. This becomes especially sensitive in contexts involving vulnerable groups, where understanding and freely giving informed consent may be limited. There is also the expectation that consent should remain meaningful even when users are not fully aware of technical infrastructures, such as networked systems or connected devices that extend data beyond the robot itself.

Empirical work highlights how these concerns surface in concrete deployments. In educational settings, researchers note that when a robot operates in small groups of children, it introduces questions about the privacy and consent of other children who are not directly interacting with it. In long-term care environments, robots may provide remote visual access in shared public spaces, raising questions about whether residents and those around them have truly agreed to being observed in this way. In schools, participants have also raised concerns about security and the possibility of robots being compromised, which could turn consent into a fragile assumption if systems are hacked and used for unintended surveillance or influence. Across these examples, consent is not only about permission at the point of use, but about maintaining trustworthy boundaries in environments where observation and data collection can easily extend beyond the intended interaction.