Chen Bing, a 63-year-old retiree who lives in Guangxi, was surprised to find her experiment with artificial intelligence tools made her the center of attention at her elementary school reunion last December.
A retired accountant with a passion for photography, Chen used ByteDance’s chatbot Doubao to divvy up expenses for the event and video generation app Dreamina to produce a photo carousel as the background for an alumni poetry reading program.
“My classmates asked me ‘How could you be so amazing?'” Chen said with a laugh, adding that she told all her friends to download the chatbot.
Since installing Doubao on her phone last year, Chen has come to rely on it for countless everyday tasks, from identifying flowers in the park to deciphering the tiny print on manuals that her failing eyesight makes difficult to read.
“I don’t need to depend on my kids most of the time,” Chen said, adding that the artificial intelligence assistant can even understand her heavily accented Cantonese.
As the most popular chatbot in China, Doubao has emerged as a frontrunner in the race to win over the country’s senior demographic. Its advertisements emphasize how accessible it can be. Seniors who are illiterate, for example, can use it easily, because there is no need to type. Users can hold natural, spoken conversations—even in local dialects—covering any topic imaginable. To foster a sense of connection, the AI even adopts familial terms of endearment, warmly addressing elderly users as “grandpa” or “grandma.”
The aging of Chinese society is accelerating and the pressure on eldercare is becoming increasingly prominent. By the end of 2025, China’s population aged 60 and above had reached 323 million, accounting for 23% of the total population. Meanwhile, the number of people aged 65 and above reached 223.65 million, making up 15.9% of the total population.
To Beijing, that is a group with big spending potential. In recent years, the Chinese government has actively promoted the so-called silver economy, estimated to account for around 10% of total gross domestic product by 2035, by launching various support policies to unlock the huge consumption potential of the elderly amid an economic downturn.
Against such a backdrop, tech companies are also competing for new users from this cohort, sometimes by offering perks. As of December, the number of generative AI users in China reached 602 million, with the penetration rate in the overall population rising to 42.8%, from 17.7% in December 2024, according to the annual report by China Internet Network Information Center, a government body.
In addition, while generative AI users are primarily younger people, adoption has begun to expand into middle-aged and older groups. Users aged 19 and below account for the largest share at 26.4%. Those aged 50–59 make up 10.1%, and users aged 60 and above account for 5.1%, according to the report.
Meanwhile, a joint survey of around 5,500 senior citizens by Alibaba and a local college for seniors last year found that though older seniors have lower AI adoption rates, once they start using AI, the group tend to be more loyal and use it more frequently.
Many of China’s colleges for the elderly now offer courses in how to use AI tools to chat, make videos and write articles.
Last year, Tencent introduced an elderly text library constructed from over 9,000 real-life question-and-answer dialogues with elderly individuals. The library is being used to train and improve its large language model’s ability to better understand the needs of senior users.
AQ, known as Ant Afu in Chinese, is the AI-driven health app launched by Ant Group in June 2025. Users can simply chat with Afu, either in text or by voice, to obtain health guidance and access relevant health services.
Slightly different from its rivals, 55% of Afu’s users come from third-tier and smaller cities, and more than 25% of its total users are aged above 55, according to the company.
Targeting services for the elderly, however, is not without pitfalls.
For example, if Afu determines that professional care is necessary, it recommends next steps, such as online consultations or hospital appointments, and suggests appropriate physicians. This has drawn concern over whether such recommendations are in the user’s best interest. Search giant Baidu used to recommend hospitals based on paid advertisements, which allowed the highest bidders to take top spots in search results regardless of medical qualification. This practice was implicated in the death of a young man named Wei Zexi in 2016 that sparked national outrage at the company’s practice as well as new restrictions on such advertising.
Amid similar doubts, Afu issued a statement in late December stating its answers “contain absolutely no advertisements or product recommendations” and that there is no commercial ranking, with answers “completely unaffected by any other commercial factors.”
Misdiagnosis by AI chatbots, meanwhile, is a global issue. A recent study by Massachusetts-based Mass General Brigham healthcare system found, through examining models by OpenAI, Anthropic, Google, xAI, and DeepSeek, that misdiagnosis could reach over 80% when patient data is incomplete.
Apart from health issues, emotional support turns out to be another selling point for these chatbots. Yu Shumin, a 78-year-old retired factory worker in China, was ill in bed last month and began chatting with Doubao about her symptoms on her phone; the chatbot immediately concluded that she was suffering from rheumatism. Surprised, she went on to share with Doubao her childhood memories of being forced to work barefoot in wet fields during winter.
“I feel so sorry for you, auntie. You were just a kid back then,” the chatbot told Yu, who was born into a poor family in the countryside of Liaoning and educated only through elementary school. The words, Yu said, made her feel seen and heard for the first time in a long while.
Despite the convenience of AI, the potential risks brought by the chatbots should not be overlooked, said Guo Tao, deputy director of China Electronic Commerce Expert Service Center, as some seniors are prone to leaking sensitive information such as ID numbers and bank card details to the chatbots, or be misled by fake medical information generated by AI.
“Over time, excessive reliance on AI for decision-making in areas such as wealth management and health could diminish their ability to think independently and critically, leaving them more vulnerable to financial scams and health-related misinformation,” Guo said.
This article first appeared on Nikkei Asia. It has been republished here as part of 36Kr’s ongoing partnership with Nikkei.