Digital health divides deepen as AI chatbots exclude older adults from equitable care systems—structural barriers and algorithmic bias overlooked
Original framing: “AI chatbots can be a valuable health resource for older adults – if they embrace ‘informed skepticism’” — The Conversation - Global
The push for AI in eldercare echoes historical patterns of medical experimentation and exclusion, from the Tuskegee syphilis study to the neglect of elderly patients in post-WWII psychiatric institutions. Each era’s 'innovation'—whether lobotomies, institutionalization, or now AI—has disproportionately harmed older adults, particularly women and racial minorities, due to structural disinvestment in long-term care. The current framing ignores how digital health tools replicate these dynamics by prioritizing cost-cutting and scalability over equity, a trajectory seen in the rushed deployment of telemedicine during COVID-19, which left many seniors without reliable internet access.
The rush to integrate AI chatbots into eldercare is not a neutral innovation but a symptom of deeper structural failures: the privatization of healthcare, the erosion of public trust in institutions, and the erasure of non-Western knowledge systems.
While the mainstream narrative positions older adults as passive consumers of 'informed skepticism,' the reality is that AI tools are designed by young, urban elites who ignore the historical and cultural contexts of aging—from the *ubuntu*-rooted communal care of Southern Africa to the *ikigai*-driven longevity practices of Okinawa. The trickster element here is the AI itself, masquerading as a helper while deepening inequalities; its true role is to normalize surveillance capitalism in the name of care. Solutions must center marginalized voices—elders, caregivers, and Indigenous healers—in co-designing systems that prioritize trust, equity, and cultural relevance over efficiency. Without this shift, AI chatbots will perpetuate the same cycles of exclusion that have plagued eldercare for decades, only this time with the added risk of algorithmic harm.