![[ Visit Client Website ]](images/banner.gif)
Methods and Results: In this symposium, researchers from UC Berkeley, the University of Washington, and the University of Minnesota will showcase ways that AI presents serious risks to older adults’ safety, privacy, and financial security. In study 1, researchers from the Aging and AI Harms Project use existing algorithmic harm taxonomies to code incidents recorded in three international AI-incident databases, specifically labeling harms affecting older adults. They discovered diverse and complex harm patterns ranging from accidental death, sexual assault, and suicide, as well as harms extending beyond the original taxonomies like extreme dehydration, falls and broken bones, going without meals, and being shut in. In study 2, researchers went further to identify the role of ageism in the AI incidents experienced by older adults. Results show significant algorithmic biases and ageism across a range of AI applications, including in AI-generated scams, staffing allocation systems, and automated decision making systems. Ageism in these AI systems resulted in premature death and loss of coverage for health and long-term care services.
Expanding on these themes, study 3 highlights how Medicare Advantage plans frequently use AI decision tools to deny healthcare services, and plan participants rarely appeal. Researchers investigated the feasibility of using an AI tool to counter service denials and assist with insurance appeals in nine clinics across three states. Results from this mixed methods study present a use case for how AI can be harnessed to combat errors produced by other AI.
Study 4 focuses on AI-enabled scams and how deepfake technologies are being used to deceive older adults, resulting in profound financial and psychological harm. The study involves an evaluation of clinician-facilitated peer support groups for victims of cyberfraud, many of whom reported that AI was used to make the scam more believable. Findings suggest that peer support groups significantly reduce shame, promote social connection, and increase scam awareness among victim-survivors, reducing their risk of future victimization.
Conclusions and Implications: Regulation lags behind the growth of AI technologies that play a significant role in the lives of older adults, whether they consent to these technologies or not. Social work practitioners must be aware of how AI can negatively impact older clients, from ageist depictions of older people to healthcare denials and deepfake scams. Tailored services are needed to address the psychosocial and health impacts of harmful AI use, and social workers need training on strategies to protect their clients.