Artificial Intelligence (AI) presents both benefits and drawbacks for people with disabilities. It offers significant opportunities to enhance autonomy, accessibility, and participation, while at the same time it poses substantial risks that may reinforce existing barriers, create new forms of exclusion, or exacerbate vulnerabilities. AI can support people with disabilities through improving medical research and creating assistive tools and consumer products that can enhance autonomy and inclusion. At the same time, risks can emerge from noninclusive design and biased datasets. From the legal and ethical point of view, accessibility, nondiscrimination, autonomy, and data protection are key concerns. To address these key concerns, AI should be inclusive “by design,” supported by fundamental rights, impact assessments, and by the cooperation of people with disabilities in development and governance processes.