日本长期的自动化尝试，即便有大量宣传与公共资金，结果仍参差不齐。在一间机构里，已停产的Softbank Pepper仍有身体在场，主要被用作与人类辅助员并排带动的运动助教；有些住户则偏好可触摸、会有反应声音的毛绒机器玩具。James Wright在其2023年《Robots Won’t Save Japan》指出，早期机器人推进常形成预期外负担，并可能错配支出，但AI浪潮后关注度再度升高。根据2025年9月至10月、覆盖21国的Ipsos民调，日本在“非常兴奋”使用AI伴侣的人数比例最低，且有46%受访者表示过去12个月没有使用AI工具。
AI照护用途正在亚洲扩散：韩国已分发14,000个以ChatGPT为核心的Hyodol给长者，中国退休者使用Doubao，腾讯为Yuanbao聊天机器人举办了超过200场工作坊。北京将银发经济视为成长引擎，引用到2035年可能达到4兆美元的预测。到2050年，几乎所有OECD国家都将成为“超高龄社会”，即65岁以上人口超过20%；日本的做法可能预示未来全球的照护政策。研究表明，机器人采用可降低离职率并提升照护品质，因为将员工转向更具人际接觡、同理与灵巧度的任务；但照护人员工资仅略高于最低工资，政策若不先提高薪资与地位，机器人只会成为替代真实照护责任的借口而非支援。
AI companion robots in nursing homes can look dystopian, but Japan’s push is driven by aging pressure: the country is expected to face a shortage of 570,000 caregivers by 2040, with limited family and community care capacity. The observations in Tokyo—including plush robots and conversational dolls—show both emotional resistance (a 97-year-old grandfather rejected cheerful AI voice chat) and practical urgency. AI is therefore being treated as a policy expedient in a rapidly aging society, not as the final replacement for human care.
Japan’s long-running automation attempts, despite major hype and public funding, have produced mixed outcomes. In one facility, a discontinued Softbank Pepper was present in body but used mainly as an exercise assistant beside a human handler, while some residents preferred a plush toy robot that responds to touch. James Wright’s 2023 book *Robots Won’t Save Japan* argued that earlier robot pushes often created unintended burdens and may have spent resources inefficiently, though AI has reignited attention. In an Ipsos poll of 21 countries (Sept-Oct 2025), Japan had the lowest share “extremely excited” about AI companions, and 46% reported no AI-tool use in the past 12 months.
Adoption is spreading across Asia: South Korea has distributed 14,000 ChatGPT-based Hyodol devices to seniors, Chinese retirees are using Doubao, and Tencent has run more than 200 workshops on the Yuanbao chatbot. Beijing frames the silver economy as a growth engine, with a cited possible size of $4 trillion by 2035. By 2050, almost all OECD countries are expected to be super-aged, with over 20% of their population above age 65; Japan’s path may foreshadow global long-term care pressures. Research also links robot adoption to lower quit rates and better care quality by shifting staff toward tasks that need human touch, empathy, and dexterity. Yet care worker pay is only slightly above minimum, so policy should first raise wages and status; otherwise robots risk becoming a way to defer care obligations instead of supporting them.