Large AI models, terminals transform eldercare

Zhang Weirong, 76, ordered food by speaking to his desktop device supported by an AI agent this morning as usual, with the voice request transmitted to the community canteens, allowing him to pick up prepared dishes or arrange door-to-door delivery.

This everyday scene at Yuexiu road in Hexi district of Tianjin exemplifies a flagship artificial intelligence senior care initiative jointly developed by Baidu Intelligent Cloud and local elderly care operator Zhongmin Jukang.

With Baidu's Wenxin large language model identifying seniors' voice requests and generating service orders within the community senior care network, elderly residents can gain well-rounded support from volunteers, community doctors and other supporting service facilities.

Yuexiu road alone is home to over 22,000 seniors, with 30 percent equipped with the You'aixiaoxin desktop care robot and regularly accessing convenient services through AI terminals.

China is witnessing unprecedented demographic pressure from population aging. Official data from the National Bureau of Statistics show the country had 323.38 million people aged 60 and above by the end of last year, accounting for 23 percent of the total population, with the elderly dependency ratio climbing steadily.

The national "90-7-3" eldercare framework reveals that over 90 percent of seniors nationwide choose home — and community-based care, while institutional nursing only serves a tiny fraction of aging residents. Severe shortages of professional caregivers have become a nationwide bottleneck.

Local figures echo the national trend. Official data from the Tianjin Civil Affairs Bureau records 3.6 million seniors across the city, yet less than 1 percent reside in professional nursing institutions, with over 99 percent relying on home and community support.

Before being transformed by AI, conventional home care relied on one caregiver coordinating dozens of seniors, bringing high communication and dispatching costs with limited service coverage. A single caretaker could only serve 20 to 30 seniors at most.

After integrating Wenxin LLM and multi-agent coordination systems, each caregiver can efficiently support 120 to 150 seniors, greatly lifting service efficiency and making AI an indispensable digital supplement to the eldercare system, said Zhan Hengbo, chairman of Zhongmin Jukang.

Du Peng, dean of the School of Population and Health at Renmin University of China and a leading national gerontology expert, said that voice-driven AI systems effectively resolve two core pain points of home-based eldercare nationwide — chronic workforce shortages and seniors' barriers to accessing timely support.

"AI does not replace human companionship, but it extends care capacity to 24-hour coverage, turns passive service responses into proactive risk intervention, and bridges the digital divide for elderly groups unfamiliar with complex electronic devices," Du said.

The integrated model balances technological efficiency and human-centric care, offering a replicable solution for cities facing aging pressures.

On Yuexiu road in Tianjin, when seniors submit requests for meal deliveries, grocery shopping, online medical consultations or emotional companionship via voice terminals, Wenxin automatically conducts semantic comprehension and intent recognition. Multi-agent collaboration distributes demands to corresponding service modules and pushes uniform orders to Zhongmin Jukang's operation backend, explained Zhang Hua, smart elderly care solution architect at Baidu Intelligent Cloud.

"The platform continuously accumulates seniors' daily dialogue habits, consumption preferences and health monitoring data to build exclusive digital profiles, enabling service teams to identify latent demands and shift from passive response to proactive care forecasting," Zhang added.

Beyond daily amenities, built-in millimeter-wave radars and linked smart mattresses detect potential falls, immediately alerting seniors, their family members and community emergency responders to secure home safety.

Having iterated its operating model over 12 years, the firm has evolved from single-service supply to dual-drive offline professional services plus an online AI platform, replicating the Tianjin model in Jinan of Shandong province, Hefei of Anhui province, Xiamen of Fujian province and other cities.

Technology boosts efficiency while human warmth remains the core of elderly care. Zhongmin Jukang pioneered the full-time caregiver position back in 2019, and the occupation Home-based Elderly Care Service Specialist was included in national vocational standards last year.

Statistics show service revenue of subdistrict stations surged after AI terminal deployment, with seniors spending an average of 350 yuan ($52) to 400 yuan monthly on the platform, dominated by meal orders and daily grocery purchases.

Zhang Hua said technical teams will continuously upgrade multiagent and emotion recognition algorithms to better interpret seniors' underlying needs through big data matching of care resources.

Powered by large AI models, technology acts as a considerate digital butler bridging the last mile of home care, delivering a replicable Chinese urban solution to actively address the nationwide challenge of population aging, Zhang said.

yandongjie@chinadaily.com.cn