生成式AI介入养老陪护服务的原则建构

The Construction of Principles for the Involvement of Generative AI in Elderly Care and Companionship

摘要: AI时代与老龄化时代交叠，人工智能与养老服务融合成为社会发展的必然趋势。当前我国养老陪护领域结构性缺口逐渐凸显，一系列养老问题随之萌发。生成式AI凭借技术优势为填补缺口提供技术基础，包括全天候服务能力，可填补人力、时间覆盖缺口；自然语言交互与即时回应能力，可填补情感即时性缺口；自适应生成能力，可缓解个性化成本缺口，在一定程度上推动当前养老服务的高质量发展。然而，人的把握能力与技术的作用力存在现实矛盾，导致AI在推动养老服务发展的同时面临以责任稀释风险、心理情感风险、隐私与自主性风险为主的困境。把握核心矛盾，立足高维视域，深入挖掘问题背后的本质联系与规律，总结出功能辅助优先、责任不可转移、人的自主性强化、数据最小化与隐私为核心的原则，以为当前矛盾的化解与未来两者的深度融合提供指引。

Abstract: As the era of AI overlaps with the aging society, the integration of artificial intelligence and elderly care services has become an inevitable trend in social development. Currently, structural gaps in China’s elderly care and companionship sector are becoming increasingly evident, giving rise to a series of related challenges. Generative AI, leveraging its technological advantages, provides a foundation for ad-dressing these gaps—offering round-the-clock service capabilities to fill workforce and time coverage shortfalls; natural language interaction and real-time responsiveness to address emotional immediacy needs; and adaptive generation capabilities to reduce costs associated with personalization, thereby promoting higher-quality elderly care services to some extent. However, there exists a practical contradiction between human agency and technological influence, leading AI-driven elderly care to face significant challenges such as risk of responsibility dilution, psychological and emotional risks, and threats to privacy and autonomy. By identifying core contradictions and adopting a high-level perspective, we can deeply explore the underlying connections and patterns behind these issues, and establish guiding principles: prioritizing functional assistance, non-transferability of responsibility, strengthening human autonomy, and minimizing data use with privacy protection at the core. These principles offer guidance for resolving current tensions and achieving deeper future integration between AI and elderly care services.