ABSTRACT

The growing availability of intelligent assistants (IAs) such as Siri and Alexa has prompted exploration of how such devices can assist older adults with daily care tasks. Yet, the potential convenience of IAs notwithstanding, current approaches are limited in part because they fail to provide users with enough sense of control over their data and interactions. The lack of agency is particularly problematic for carrying out health information tasks; older adults find it challenging both to understand the data returned by the IA and to determine who has access to the critical personal and health data they share. This project will advance our understanding of how to design IAs that positively influence older adults' sense of autonomy when using them to carry out health information tasks at home. To this end, the research team will leverage existing partnerships with senior organizations and informal caregivers in the Indianapolis community to ascertain specifics about their existing interactions and challenges with IAs; these findings will then be used to establish and evaluate novel design principles that can support more transparent conversational IA interactions for health information management. Project outcomes have the potential to transform IA design and development from black boxes that provide information services to more transparent conversational partners that are able to garner trusted human-AI relationships. Additional broad impact will derive from involvement and mentoring of undergraduate and graduate students, including women and minorities, in the research activities.

The work will involve two main thrusts. The first will identify the important values that shape the help-seeking beliefs of older adults in existing informal caregiver relationships and settings; based on those findings, the team will engage older adults to collaboratively design new IA interaction strategies aligned to their needs and expectations during health information tasks. In the second thrust, system prototyping and user-centered evaluation activities will identify the potential and limitations of the resulting design strategies in addressing human autonomy needs and IA acceptance among older adults, examples to include a heightened focus on providing alternative information sources and answers as a response to users' questions, and providing pathways to access privacy settings and personal data sharing preferences during interactions with the IA.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Please report errors in award information by writing to: awardsearch@nsf.gov.