Hangzhou Elderly Living Alone Get Robot Roommates in Three-Year Trial

During the Mid-Autumn Festival, Zhao Yongfu, an elderly man living alone in Yunxin Community, Sijiqing Subdistrict, Hangzhou, gained two new roommates: robots Yuanluobo and Yuanbao. A simple call is enough to prompt a quick response from either machine, giving the home a sense of having someone available to talk to.

After meals, Zhao plays several games of chess with Yuanluobo every day, a routine that has become part of his daily life. The board games are not just entertainment; they also give him a regular activity to return to.

Yuanluobo focuses on cognitive entertainment and mental exercise, using games and training to help older users stay mentally active. Yuanbao leans toward emotional companionship and everyday question-and-answer interactions, offering conversation and simple responses.

The presence of the two robots has brought activity and company to a home that was previously quiet. They cannot replace real family ties, but they offer a new form of companionship for solo living — and serve as a starting point for observing how robots fit into households.

The robots are designed for ease of use: no complex setup, mainly voice interaction. For Zhao, the robots are not a novelty but part of his routine.

At the beginning of August, a social experiment named “Intelligent Agents Serving People” launched in Hangzhou. Its aim is to explore what changes occur when robots move into the homes of elderly people living alone. The project emphasizes long-term observation rather than a one-off technology demonstration. The evaluation is intended to track changes over a meaningful period, not just initial reactions.

The project is jointly led by Zhou Chaonan, a member of the National Committee of the Chinese People's Political Consultative Conference and chairman of Runze Group, and Professor Su Jun, dean of Tsinghua University's Institute for AI Governance. It plans to place intelligent agent products in 300 to 500 households nationwide, provide supporting services, and conduct a three-year follow-up evaluation.

The project maintains a public-welfare orientation: devices and services are provided free to participating families and are continuously updated and optimized based on the elderly users' actual needs and feedback. The arrangement aims to remove cost concerns so that older people can accept the robots more naturally. Feedback loops are built into the design so products can evolve with users.

As the first stop after leaving Beijing, Hangzhou has already brought companion robots into 30 elderly households in Jianqiao, Xiaoying, and Sijiqing Subdistricts of Shangcheng District. Most of these households are elderly people living alone or “empty nest” families. The pilot is small by design, focusing on depth rather than scale.

A three-year evaluation period is unusual for robot deployment projects. Most trials of robots entering homes last from a few weeks to a few months and focus mainly on short-term acceptance and frequency of use. The three-year timeframe addresses a different set of questions. Short trials can show whether people accept a robot, but they cannot show how habits evolve.

It asks whether long-term interaction changes elderly users' habits, whether robots can adapt to changes in an older person's physical and cognitive condition, and whether the companion relationship stabilizes over time. These questions matter because aging can bring gradual changes that short pilots may miss.

The project provides devices and services free of charge, lowering the barrier for elderly people to try the technology. But it also means the project's sustainability depends on whether a suitable paid model can be found later. The question of who pays later remains open.

The 30 households represent a small share of the total elderly population living alone in Shangcheng District. The value of the pilot lies not in its coverage but in whether it can produce a replicable service process and evaluation method.

From a broader perspective, such experiments seek to answer whether robots can take on part of daily companionship and cognitive training in elderly care, and how society should evaluate this new type of relationship. If the process works, it could be replicated in more communities and households. The project is therefore as much about learning how to assess human-robot companionship as it is about deploying devices.