With the European Accessibility Act having come into effect in June 2025 and the EU AI Act the summer prior, policies around accessibility and artificial intelligence are converging. For organisations in Ireland, this means considering accessibility and inclusion from the get-go when designing and deploying AI systems.
For people with disabilities, AI offers significant opportunities to enhance accessibility and independence, while also raising new risks that demand careful oversight. A recent report by the Centre for Excellence in Universal Design at The National Disability Authority explores these dynamics in the Irish context, analysing how AI may reshape six critical sectors—education, employment, healthcare, financial services, active citizenship, and leisure.
AI technologies—particularly multimodal systems that combine text, speech, images, and video—are opening new ways for people with disabilities to communicate, learn, and engage with services. For example, Large Language Models (LLMs) can convert text to speech, simplify technical documents, or generate real-time translations. In healthcare, AI holds potential for early diagnosis and personalised treatment. In education, it can adapt learning materials to individual needs. Irish employment services could use AI to match people with disabilities with jobs more effectively, while AI-driven tools in finance may improve access to banking and benefits. Even in leisure, AI has begun to make cultural, gaming, and travel experiences more inclusive.
The thread connecting these developments is AI’s potential to remove barriers that have historically limited participation for people with disabilities. If designed using a universal design approach, AI could help build a more inclusive society where accessibility is embedded in the core of digital and public life.
Yet, alongside these opportunities lie serious risks. People with disabilities are significantly underrepresented in AI training data. This means that systems may produce biased or exclusionary outcomes, from job-screening tools to credit scoring models. Moreover, AI interfaces are designed around limited interaction modes, excluding people who rely on alternative communication methods. Even “explainable AI” tools that aim to make decisions transparent often rely heavily on visuals, leaving out users with visual impairments. Privacy and security concerns also loom large, particularly when sensitive health or disability data is involved and when poorly designed systems might deny supports or misinterpret disability-related needs.
These risks underline the importance of involving people with disabilities in AI development and ensuring universal design principles guide both technology and policy.
The aforementioned report from the Centre for Excellence in Universal Design underscores that Ireland is at a pivotal moment. The promise of AI lies in its ability to transform services and daily life for all people, including those with disabilities, but this potential will only be realised if universal design is prioritised from the outset. Effective regulation, cross-authority collaboration, and ongoing involvement of disabled people are critical to ensuring AI enhances inclusion rather than creating new barriers.
As AI moves from concept to lived reality, the choices made today—in design, regulation, and oversight—will shape whether it becomes a tool for empowerment or exclusion. As AI implementation expands, the Centre for Excellence in Universal Design will continue to share relevant information about AI, accessibility and Universal Design: universaldesign.ie/ai.