AI and Accessibility: The questions every organisation is asking

AI and Accessibility: Key Questions for Organisations

Over the past year, conversations around accessibility increasingly centre on artificial intelligence (AI). The shift from exploration to deployment is marked: many organisations are now embedding AI tools into their products, services and processes. For those working in accessibility and digital inclusion, this evolution carries significant implications.

AI’s Early Benefits for Disabled People

Robin Christopherson, a blind accessibility expert, highlights that AI has been enhancing life for disabled people well before generative AI became widely known. Early AI-powered tools have long helped with reading printed documents, recognising objects, identifying products and interpreting the physical environment. These benefits have already transformed countless everyday tasks for many disabled individuals.

Four Central Questions in Accessibility and AI

The accessibility community now focuses on four main questions:

How can AI-generated code scale accessible practice, or risk replicating existing barriers?
Where does automated accessibility testing end and human judgement begin?
How should governance and accountability around AI be structured?
What must organisations do to build and maintain trust in AI-enabled services?

These topics will be explored in detail at AbilityNet TechShare Pro on 18 and 19 November, bringing together accessibility practitioners, disabled users, technology leaders and policymakers.

Building Accessibility into AI-Assisted Development

The rapid rise of AI-assisted software development has created pressures and opportunities. Accessibility requirements, trusted design patterns and testing methods can now be embedded within AI-assisted workflows. When properly managed, AI can help teams detect barriers earlier and apply accessible components more reliably.

However, there is concern about AI models learning from existing examples, including the vast amount of inaccessible websites and confusing interfaces online. This risks scaling exclusion if not carefully controlled. Accessibility professionals need to influence AI strategy, governance and design frameworks to prevent systemic barriers from multiplying.

The Limits of Automated Testing

AI-powered accessibility tools are improving, able to identify certain issues, flag risks and offer guidance. Yet, human expertise remains essential. Accessibility depends on the real-world experience of users completing tasks, understanding information and recovering from errors. Automated checks—such as detecting alternative text or keyboard navigation—cannot replace human insight into whether user experiences are genuinely accessible.

Governance: An Accessibility Imperative

Organisational governance around AI has become critical for accessibility. As AI systems influence recruitment, customer service, content creation and more, questions of bias, transparency and accountability grow in importance. Disabled people have often been under-represented in design and testing, risking reinforcement of exclusionary practices unless safeguards are in place.

Governance needs comprehensive oversight, especially in regulated sectors like financial services, where inaccessible AI-driven decisions may affect customers’ access to credit or support. Oversight should cover the customer journey, explanation quality, access to human assistance and evidence that fair outcomes are delivered.

Building Trust in AI

Trust underpins the effective use of AI. Organisations seek the benefits of AI, customers want more accessible services, and employees desire useful tools. However, confidence depends on accessibility being prioritised and robust support being available when AI systems fail.

Mark Walker, AbilityNet’s Head of Marketing and organiser of TechShare Pro, notes the sector’s focus “on balancing risk and reward.” Accessibility professionals play a vital role in ensuring disabled people shape innovation rather than being overlooked.

Questions for Accessibility Leaders

Christopherson suggests accessibility leaders consider three practical questions:

Where is AI already in use within the organisation, including informal applications?
What evidence demonstrates AI works for diverse users, beyond automated checks—such as assistive technology testing and user research?
How much influence does accessibility have over AI strategy, governance, design and monitoring?

Sharing Knowledge and Shaping Practice

Many organisations are engaging with these questions across sectors including finance, retail, technology and public services. AbilityNet TechShare Pro provides a platform for sharing experiences, featuring sessions on AI-generated code, AI testing methods, procurement, governance and risk management.

The event is open both in London and online, with free online tickets for AbilityNet’s tenth anniversary. It aims to translate rapid technological developments into informed, inclusive practices.

The Principles of Inclusive AI

Despite fast-evolving tools and questions, the core principles remain constant: involve disabled people and diverse users; gather reliable evidence; maintain clear accountability; and test real user experience. Following these habits increases AI’s potential to broaden access and independence.

How AbilityNet Can Support Organisations

Many organisations now face questions about how accessibility teams should engage with AI, develop governance and testing strategies, and ensure disabled customers and colleagues are included as AI adoption grows.

AbilityNet offers expertise through accessibility audits, user testing with disabled people, training, consultancy and AI-focused accessibility guidance. More information is available about AbilityNet’s accessible AI integration and consultancy services. Organisations can also contact AbilityNet’s experts to discuss assessing AI features, customer journeys or deployments.