Globally, as societies become both older and more intelligent, ensuring the meaningful participation of older adults in the AI-driven world has become a critical concern for policymakers and researchers. Yet, research at the intersection of ageing, AI, and digital inequality remains fragmented and largely Western-centric, with limited empirical evidence from rapidly developing contexts such as China. China’s transition creates a dual landscape of opportunity and risk for older adults. This “silver digital divide” reflects inequities in cognition, access, usage, and evaluation that heighten exclusion in the intelligent era. The intergenerational knowledge gap prevents many older adults from engaging in online interaction or civic discourse, rendering them “information poor”. This study adopts an interdisciplinary, mixed-qualitative design to explore how older adults can be meaningfully integrated into the AI-driven society. First, drawing upon an extensive literature review and policy analysis, this research maps both the global and Chinese landscapes of AI adoption among older populations, identifying multiple dimensions of digital inequality. The Cognition–Access–Usage–Evaluation (CAUE) framework is applied to systematically categorise the multi-layered digital divides experienced by older adults in the AI context. Empirical evidence is then generated through semi-structured interviews with senior citizens residing in both urban and rural settings, along with their adult children. Thematic analysis reveals that intergenerational technological support—encompassing emotional, informational, and instrumental forms—plays a pivotal mediating role in enhancing older adults’ cognitive readiness, physical and networked access, operational competence, and evaluative capacity. To strengthen analytical rigour and inclusivity, these qualitative findings are triangulated with a supplementary survey measuring AI literacy and attitudes toward AI adoption. However, structural constraints such as algorithmic bias and unequal resources persist. This study concludes by advocating coordinated action across government, industry, and families to build equitable, age-friendly AI ecosystems.