A sociotechnical approach to ethical AI in disability services: insights from stakeholders

The integration of Artificial Intelligence (AI) in social services is transforming how these services are delivered, and is prompting questions about its responsible use. This paper draws on 30 semi-structured interviews with disability service providers, technology developers, academics, and disabled people to examine their relationships with emerging AI technologies. Our findings show that principal inhibitors to effective AI integration into disability services lie predominantly in the social, rather than the technical, domain. We conclude that careful planning, co-design of digital infrastructure with persons with lived experiences of disabilities, and intentional multi-stakeholder collaboration are essential for safe and equitable adoption of AI technologies. Capacity for deeper transformation also lies within organisational agency: disability service providers hold the power to move beyond top-down rollouts of AI toward a human-centered approach that empowers frontline workers and disabled people, and enables their meaningful participation in this technological shift.