Privacy – Profiling
The robot collects data continuously or aggressively; data safety and protection.
Profiling in human-robot interaction refers to concerns around the extensive collection, aggregation, and use of personal data by robotic systems. As robots move from controlled lab environments into homes, schools, and care institutions, they increasingly become data-intensive systems that observe, record, and infer aspects of users' lives. This creates a tension between functionality and surveillance, where the same sensing capabilities that enable assistance can also produce detailed and potentially sensitive profiles of individuals.
This topic is strongly tied to the broader value of privacy, particularly in how robots collect, store, and transmit data over time. Profiling concerns focus on the risks of continuous and sometimes aggressive data collection, which can include behavioural patterns, emotional states, routines, and interactions. Such data may be misused by companies or accessed by malicious actors, but concerns also arise in more ordinary scenarios: for example, data from care settings being shared with insurers or third parties without clear benefit to the user. The risks are especially pronounced for vulnerable groups such as children or older adults, where consent may be mediated by caregivers. Context also plays a central role: in-home deployments can increase exposure to misuse due to weaker oversight, including in situations involving domestic abuse. Across studies, participants also emphasise that privacy protection must persist even when users are not aware of, or not able to manage, the full extent of data collection happening around them.
Across the literature, profiling is often expressed through unease about being continuously observed and the uncertainty of future data use. In education contexts, concerns appear as "the unease of not knowing what could happen with all the data collected by robots in the future".
In educational and care settings, profiling becomes particularly sensitive when robots collect emotional or physiological data. For instance, a study notes that "physiological measures, and emotional facial expression recognition, offer the potential to detect and possibly record information about the emotional state of children interacting with a robot", raising questions about the depth of inference possible from seemingly simple interactions. Furthermore, in another work, concerns are framed around "the user's privacy which could be compromised by creating a sensitive consumer profile", showing how long-term interaction can lead to detailed behavioural categorisation. Even systems designed with minimal storage illustrate how data transmission and temporary storage still raise meaningful privacy concerns.
Taken together, these perspectives show profiling as a persistent and practical issue in human-robot interaction. It is not limited to data security alone, but extends to autonomy, consent, and the social meaning of being observed.
Excerpts from the paper:
About the value "Privacy"
The value of privacy is the one encompassing the largest amount of topics as highlighted in both the scoping review and the focus groups. It includes – among others – the right to be informed, to access and share the data collected, and the issues related to being continuously under the watchful eye of a robot. The focus groups participants have also highlighted that this value is strongly related to where, how and in which context the robot is used. Using a robot at home is completely different from using it in a public space. The experts emphasised that privacy should be maintained even when users do not want others to know they are using the robot.
About "Profiling"
This topic covers all those concerns related to the excessive quantities of data collected by robotic devices. The continuous and aggressive collection of personal data for profiling risks misuse by both collecting companies and malicious actors. The focus groups participants expressed concerns that data from elderly care centres could be shared with insurance companies or sold to third parties without benefitting the user. These concerns are heightened for vulnerable groups like children, where parents or guardians should manage data consent. Additionally, the context of the application, such as in-home settings with less stringent regulations, can increase potential risks and harms, such as misuse by domestic abusers. Other concerns, emerged during the scoping review, touch on the safe transmission and storage of this data.