Artificial intelligence is already shaping how we communicate, make decisions, and whose lives count. Across culture, it quietly designs the rules we live by – sorting, predicting, filtering, and responding. Dominant narratives frame AI either as a neutral tool or an existential threat, but both positions obscure a more urgent question: who benefits from the worlds AI helps to build, and who is rendered expendable in the process? For disabled people, AI is rarely abstract. It is already entangled with survival, access, and participation.
Disabled people are using AI to interpret Whaikaha Ministry of Disabled People and Ministry of Social Development communications, draft funding appeals, translate policy into plain language, and manage the cognitive labour required to navigate increasingly restrictive systems. In an era of tightened eligibility, surveillance, and administrative burden, AI redistributes time, energy, and interpretive power. These uses are not about optimisation or convenience; they are about endurance. Within disability-led arts practice, technology is similarly not a prosthesis that compensates for lack, but a collaborator that reshapes how agency circulates.
The question, then, is not whether AI will change society – it already has – but who gets to author the worlds that emerge. Disabled people have always been world-builders, often under conditions of constraint and necessity. Participation has never been isolated or self-sufficient; it has always been collective, mediated through supports, technologies, environments, and relationships. AI simply makes this interdependence visible. Rather than asking how AI can increase independence or efficiency, we might instead ask what becomes possible when interdependence is treated as the baseline rather than the exception.
Disability is still widely framed as a problem to fix or a cost to manage. Yet disability is also a site of knowledge. Fluctuating capacity, assisted communication, supported decision-making, and collective care are not failures of the system – they are how social life actually functions, even when it refuses to admit it. From this perspective, disability offers a critical lens for thinking about AI not as a replacement for human agency, but as a mechanism for redistributing it: slowing it down, spreading it out, making it accountable.
This politics of redistribution is explored in Abilitopia, a disability-led dance-theatre work by Touch Compass premiering at Te Pou Theatre (26–28 February). In the work, human dancers perform alongside a fully AI-enabled robot, asking whose values shape technological worlds and whose realities are designed out. Rather than projecting a speculative future in which technology finally includes disabled people, Abilitopia starts from the present. Disabled people are already co-authoring technology every day through adaptive systems, assisted communication, and collective problem-solving. The work asks what might change if these practices were recognised not as accommodation, but as forms of expertise.
As a disability-led production, Abilitopia does not simply represent disability on stage; disability structures the world of the work itself. Agency is shared, visible, and negotiated across bodies, technologies, sound, light, and space. Dependency is not hidden or overcome – it is made central. In doing so, the work quietly destabilises a core cultural assumption: why is dependence treated as failure in humans, but as a marker of sophistication in machines? The robot does not dominate or replace the dancers; instead, it exposes how all agency is relational, contingent, and co-produced.
AI unsettles us because it performs actions culturally coded as human – speaking, responding, interpreting—while lacking consciousness or embodiment. Its power lies not in personhood, but in effect. When viewed through a disability lens, that effect becomes political. AI reveals how much of what we call “independence” is already scaffolded, assisted, and distributed, even as disabled people are routinely punished for making those supports visible.
The worlds disability practices gesture toward are not hypothetical. They already exist in Aotearoa. Many AAC users customise communication systems to reflect Te Ao Māori understandings of collective voice, whakapapa, and relational identity. When AI prediction enters these systems, voice becomes co-authored – neither fully human nor machine, but relational. Capacity is assumed to vary. Participation does not require speed, autonomy, or fluency in dominant modes of expression. Agency is collective, revisable, and shared.
This is where disability fundamentally reframes the conversation around AI. It challenges the fantasy of frictionless autonomy and replaces it with an ethics of relation. It insists that the ways disabled people build worlds under constraint – slowly, collaboratively, improvisationally – are not marginal adaptations to be tolerated, but vital models for our cultural future. If AI is already shaping the worlds we inhabit, then disability offers not a warning, but a blueprint for how to live there together.
