Balancing personalization and privacy

The promise of Socially Assistive Robots (SARs) to deliver personalized care for older persons depends heavily on extensive data collection, raising concerns about privacy, autonomy, trust, and the limited involvement of older adults in shaping data-related configurations. This study examines how developers and researchers navigate the tension between personalization and privacy in the datafication process in SAR development for older people. We conducted eighteen interviews with international experts from industry and research institutes and analyzed them using a constructivist grounded theory approach. Participants identified 41 unique data types used by SARs, comprising numerous variables. In describing the datafication of older persons as SARs users, participants characterized SARs as capable of ‘collecting everything.’ Regulatory compliance and data safeguarding were framed in ‘all or nothing’ terms, leaving older adults uninvolved in personalizing or influencing data management and control over their data, effectively requiring them to accept the SAR as a whole or decline it entirely. At the same time, participants viewed control, transparency, personalization, and the storage of personal data for user identification as essential for effective social robotics. These findings reveal a personalization-data-storing paradox: privacy preserving practices and regulatory constraints often restrict data use and online connectivity, yet meaningful personalization and interaction depend on richer data use and access. We illustrate this paradox through a typology of four SAR categories and highlight the need to involve older persons in shaping meaningful data management practices that balance privacy protection with an enriched user experience

Personalization · Qualitative research · Safeguarding · Typology · AI in Service Interactions · Social Robot Interaction and HRI · Technology Use by Older Adults