This paper provides a critical, historically situated investigation of policy discourses surrounding the digital inclusion of older people over the past quarter of a century – from the e-society of the early 2000s to today’s AI policy visions. Drawing on data from the AGEAI research project ( www.anyage.ai ), we investigate emerging debates on AI literacy in Europe, focusing on the role of digital skills in the AI-driven economies of the future. Specifically, we examine the AI strategies of four European countries: Germany, Spain, Poland and the UK, through textual analysis. Our focus is on older people, a growing social demographic mostly associated with healthcare and social welfare solutions. We situate current iterations of AI literacy in their historical-technological context, spanning over two decades of media literacy and digital inclusion policies. Our analysis reveals that AI skills are predominantly conceptualised as user capabilities, and ‘users’ are addressed as members of homogenous groups, echoing earlier policy narratives in which media literacies were framed in binary terms. Individuals either ‘have’ or ‘have not’ developed their digital media skills and capabilities for internet access and use. Older people are frequently characterised using ageist tropes, such as deficiencies in learning, and vulnerabilities to technological exposure. At the same time, they are positioned at the centre of an expanding market for technological solutions. We identify a new set of challenges for research and policy, arising from the fact that AI systems do not always require the active involvement of the ‘users’ on whom they are deployed and can further exacerbate social and digital inequalities related to age.