Instead of handing older adults a finished robot and asking them to react, researchers at the University of Houston gave them air-dry clay and asked them to build one. The exercise, run at Richmond’s Mamie George Community Center, is part of a broader effort by the university’s Empathetic Lifespan AI & Robotics for Aging Lab — ELARA for short — to design AI companion robots meant to ease loneliness among older adults without talking down to them or treating them as passive test subjects.
Reframing Aging as a Design Problem, Not a Tech Problem
ELARA is directed by Chorong Park, an associate professor at UH’s Gerald D. Hines College of Architecture & Design, who argues that the mental health and isolation risks facing older adults are best addressed through better design rather than more powerful hardware. Park pushes back directly on what she calls “techno-ageism” — the assumption baked into much of the industry that older people simply can’t learn new technology. “Technology designs are based on assumptions older people can’t use them, but that’s not true,” Park says. In her view, the failure sits with designers who don’t explain or build for this population, not with the users themselves.
Treating Older Adults as Co-Designers, Not Subjects
The clay-modeling sessions were led by undergraduate researchers Anika Vadlamudi and Jana Qaddoura, who had participants evaluate robots already on the market — including Emo, Moflin, and Cupboo — before sculpting their own ideal companion by hand. The goal was to surface preferences that standard surveys or usability tests tend to miss. A parallel study on UH’s campus gathered reactions from younger adults, and the overlap in results was notable: both age groups gravitated toward practical functionality, like built-in timers, medication reminders, and weather updates, over flashier conversational features.
The Scale of the Loneliness Problem
The stakes behind the research are well-documented. A report from the National Academies of Sciences, Engineering, and Medicine found that nearly 25% of adults age 65 and older experience social isolation, a condition linked to elevated risk of dementia, heart disease, and stroke. That statistic has driven a wave of commercial entrants into the companion-robot space — from Intuition Robotics’ ElliQ, already deployed in apartments across New York City, to South Korea’s Hyodol, which is preparing a 2026 US launch with English, Chinese, and Japanese language support, and Ryan, a humanoid companion tested with six older adults with early-stage dementia or depression who reported feeling happier after weeks of round-the-clock access.
The Surveillance Trap That Kills Adoption
Park’s own peer-reviewed research, published through the ACM, found a specific and often-overlooked barrier to adoption: fear of “severe surveillance.” Older adults who feel constantly monitored by a device meant to keep them company tend to reject it outright, no matter how sophisticated its conversational abilities are. “They won’t use the robot if they don’t feel at ease or constantly watched,” Park says, a finding that puts her lab at odds with companion-robot designs built around continuous health monitoring and data collection as a selling point.
A Real Risk for Users With Dementia
Not every voice in the field is optimistic that AI companions are an unambiguous good. Researchers studying the technology’s use among people with dementia have flagged a specific danger: users vulnerable to delusions or confusion may have those states reinforced rather than gently corrected by an AI companion designed to be agreeable and validating. That risk sits uneasily alongside the loneliness-reduction benefits the technology is meant to deliver, and it underscores why Park’s team insists design choices — not just algorithmic sophistication — determine whether a companion robot helps or harms a vulnerable user.
Scaling Beyond a Single Community Center
ELARA’s work sits inside a much larger institutional push: the University of Houston currently leads roughly $70 million in AI-related research spanning more than 140 faculty researchers. Park is now working to extend her lab’s co-design methodology into an international research network, arguing that insights from one country’s older population can inform product design elsewhere. “Aging is a global challenge, and we have things to learn from each other,” she says.
What’s Next for Companion Robot Design
The clay models themselves will never end up in a factory, but Park’s team is feeding the underlying preferences — for privacy, for simple practical utility over novelty, for a sense of control rather than surveillance — into design guidelines meant to shape the next generation of aging-in-place technology. Whether the well-funded companies already racing into this market, from Intuition Robotics to Hyodol, actually adopt those principles is an open question. For Park, the answer will determine whether AI companions become a genuine tool against isolation or one more piece of technology built for older adults instead of with them.
