Abstract
Over the last decade, bioethics has begun to address the ethical issues emerging as artificial intelligence (AI) and associated technological processes such as automated decision-making (ADM) become part of healthcare and research. Recent work on justice in AI demonstrates that supposedly neutral AI systems can perpetuate the marginalization of various communities. But so far, there has been little exploration of the interaction of AI and disability. In this empirically based project, we have explored the implications of ADM in the lives of people with disability in Australia. This paper focuses on a point that was consistently raised in discussion by disabled participants but is rarely encountered in the AI ethics literature, especially in relation to disability: the problem of automated systems' failures of recognition.