Protecting Privacy of Persons with Dementia (2026-2027)

Background 

AI companion technologies, such as conversational chatbots and smart speakers, are increasingly being used as supportive tools for dementia care. These systems offer potential benefits, including emotional support, reminders and social engagement for persons living with dementia and their care partners. Yet, unlike health technologies governed by HIPAA, commercial AI companions operate with little regulation or oversight. 

These tools often rely on anthropomorphic design that encourages users to disclose intimate details, including health information, financial concerns and personal history. Prior research shows that both persons living with dementia and their care partners lack access to clear, comprehensible explanations of how these systems collect, store or use their data. Privacy lapses can worsen psychological stress, intensify relational tensions and, in some cases, contribute to caregiver abuse. Despite these risks, no existing policies meaningfully address the privacy needs of this vulnerable population. 

As advanced AI technologies become more integrated into care settings, there is an urgent need to safeguard privacy, reduce harm and ensure equitable, informed use of AI companions for dementia care. 

Project Description 

This project team will examine privacy risks associated with commercially available AI companion tools used by persons living with dementia and their care partners. The project includes four major phases: 

Preparation 
Team members will receive training in research ethics, dementia care contexts, and qualitative and quantitative user research methods, including surveys and Delphi studies. 

Data collection 
Work will begin with a Systematization of Knowledge (SoK) review of existing security, human–computer interaction (HCI) and health research to identify gaps in literature. The team will then conduct a comprehensive audit of AI companion apps and devices, analyzing privacy disclosures, interface design and data flows. 

This work will be complemented by semi-structured interviews with privacy experts in academia and industry as well as an online survey of approximately 300 care partners and semi-structured interviews with eight to 10 persons with early-stage dementia to explore their privacy needs and expectations. 

Data analysis 
Quantitative survey data will be analyzed with parametric and nonparametric methods. Qualitative data will undergo thematic analysis using two cycles of coding. Frameworks such as Contextual Integrity will guide the interpretation of major themes. 

Dissemination and next steps 
Findings will inform the design of user-centered privacy protection tools for persons living with dementia and their care partners. The team will prepare for future participatory design workshops, accessibility testing, academic publications and public-facing workshops to amplify findings and build cross-sector partnerships. 

Anticipated Outputs 

• A stakeholder-informed framework of privacy risks and solutions for AI companions in dementia care
• An open-source privacy dashboard offering protective tools and resources for patients and caregivers
• A grant proposal for scaling the project
• Privacy education materials and advocacy events for dementia care communities, policymakers and other key stakeholders 

Student Opportunities 

Ideally, this team will include 2 graduate students and 4 undergraduate students working across technical and human-centered domains. Participants may come from computer science, engineering, psychology, public policy, population health, nursing, social sciences and related fields. Applicants should be passionate about community-engaged research that advances data privacy and promotes the well-being of aging populations with dementia.  

Students will gain skills in: 

• Mixed-methods research, including SoK reviews, audits, surveys and interviews
• Qualitative and quantitative data analysis
• Prototype development of privacy-protective tools and user interfaces
• Human-centered design principles, accessibility considerations and research ethics in dementia care
• Engagement with community partners and end-users in participatory research
• Communication of findings through presentations and academic publications 

Graduate students benefit from the experience of leading subteams and mentoring undergraduates. One graduate student will be selected to serve as project manager. 

Timing 

Summer 2026 – Summer 2027 

Summer 2026 (required): 

• Perform foundational SoK research and initial data collection 

Fall 2026:

• Continue data collection and begin data analysis

Spring 2027

• Complete data analysis, develop interventions and disseminate findings 

Summer 2027 (optional): 

• Refine interventions, prepare publications and conduct additional outreach 

Crediting 

Academic credit available for fall and spring semesters