Truace
Subscribe

Being Sensed, Scored, and 'Connected': A Social-Gerontological Critique of AI-Driven Loneliness Detection for Older Adults in Taiwan

Enthusiasm for artificial intelligence (AI) that continuously infers emotion promises to change how care systems identify older adults at risk of loneliness. We use AI-driven loneliness detection to mean systems that infer loneliness, or a proxy for it, from sensed data rather than from a person's answer, rendering the inference as a score or category used in care; they range from facial and vocal emotion recognition to behavioral risk prediction from everyday activity data. Engineering feasibility, more than an…

By the Truvace desk2 min read1 sourcePeer-reviewedTRV-2026-1774· cite & verify
Being Sensed, Scored, and 'Connected': A Social-Gerontological Critique of AI-Driven Loneliness Detection for Older Adults in Taiwan
Field test and evaluation of a functional assessment system for adults needing long-term care : executive summary by Errecart, Michael T Strope, Donald H RMC Research Corporation United States. Social and Rehabilitation Service. Public domain

In brief

A Forum article in The Gerontologist examines AI systems that infer loneliness, or a proxy for it, from sensed data such as facial and vocal emotion recognition or everyday activity patterns, and turn the inference into a score used in care. It focuses on Taiwan's ageing society and long-term care system.

The authors write before any routine use, so the concerns are anticipatory rather than measured harms. The article matters because it argues that design should start from older adults' own meanings and expectations of privacy and autonomy, and it leaves open how older adults will actually negotiate being sensed and scored.

Main points

  1. The article defines AI-driven loneliness detection as inference from sensed data rather than a person's own answer, rendered as a score or category used in care.
  2. It says engineering feasibility, more than social-gerontological understanding of loneliness, has driven this agenda.
  3. It names three tensions: loneliness reframed as a biometric deficit, emotional restraint read as noise, and care slipping into surveillance.
  4. It proposes principles for a socially accountable approach and a research agenda.

The problem

Before any routine deployment in Taiwan, AI loneliness detection could recast a relational experience as an individual biometric deficit, misread culturally patterned emotional restraint as noise, and slide from care toward surveillance.

The rundown

The article states that such detection is not yet in routine use in Taiwan. It therefore takes an anticipatory-governance stance, examining before deployment a trajectory that policy makes plausible.

The technologies covered range from facial and vocal emotion recognition to behavioral risk prediction from everyday activity data, all producing scores or categories for use in care workflows.

The authors take seriously the case for technological augmentation of scarce care resources, but argue that culturally responsive design should begin with older adults' relational obligations and expectations rather than with the sensors.

Sources

  1. Peer-reviewedThe Gerontologist2026-10-07

The debate