Abstract Rapid population aging is intensifying the mismatch between the rising demand for eldercare and the shrinking care workforce. Voice-based artificial intelligence (AI) companions—such as smart speakers and conversational agents—are increasingly deployed to address social isolation, a critical but underrecognized threat to aging-in-place. While these systems offer scalable emotional support, personalized interaction, and privacy-preserving monitoring, they also introduce ethical risks including artificial empathy, emotional dependency, and algorithmic age bias. This commentary argues that voice AI companionship should be governed as a risk-tiered, dignity-centered triage-and-amplification layer within human care—not as a surrogate relationship. We outline a governance framework emphasizing human oversight, dignity-by-design, risk-tiered regulation, and evidence requirements for reimbursement. Harnessed responsibly, voice AI can strengthen independence and safety for older adults; deployed without safeguards, it risks replacing authentic connection with a superficial simulation of care.
Kuan-Ting Kuo (Wed,) studied this question.