Ethics surrounding artificial intelligence-powered assistive technologies used by people with visual and hearing impairments: a systematic review

This paper explores the integration of artificial intelligence into assistive technologies (AI-ATs) for individuals with visual and hearing impairments. While these technologies enhance independence and quality of life, they raise complex ethical challenges demanding systematic examination. We conducted a systematic review synthesising findings from 81 peer-reviewed publications (2010-2024), searching nine databases with rigorous inclusion criteria. An analytical framework examined 11 ethical dimensions across various implementation contexts. The analysis reveals equity and accessibility as primary ethical concerns, followed by autonomy and privacy protection. Workplace applications dominate the literature, while educational contexts remain underrepresented. Implementation strategies prioritise user-centred design but show limited attention to technical accountability. Findings indicate critical misalignment between ethical principles and practical implementation. This review contributes a comprehensive framework for evaluating AI-AT ethics and highlights the need for integrated development approaches emphasising interdisciplinary collaboration among technologists, ethicists and end-users.