Over the past decade, global policy discourse has increasingly focused on expanding “internet access” as a cornerstone of digital inclusion. Yet for persons with disabilities, connectivity alone does not translate into meaningful participation. A user may be connected, possess a device, and have basic digital skills—and still remain excluded due to inaccessible design, incompatible platforms, or poorly implemented technologies. This gap calls for a conceptual shift: from access to agency. At the same time, the rapid integration of artificial intelligence into digital ecosystems is reshaping how users interact with technology. AI-powered tools—such as automated captioning, voice interfaces, and assistive agents—hold transformative potential. However, they also introduce new risks, including bias, unreliable outputs, and inaccessible design layers that may deepen exclusion if left unaddressed. Against this backdrop, accessibility must be reframed not as a technical afterthought, but as a governance, rights, and development issue—central to debates on AI, digital public infrastructure, and inclusive innovation.