Accessible Intelligence

Accessible Intelligence

Making AI-generated intelligence accessible, understandable and actionable for diverse people.

Accessible Intelligence is an integrative perspective that connects research in accessibility, human-centred AI, explainable AI and responsible AI. It foregrounds whether people with diverse abilities, knowledge, languages and circumstances can meaningfully access, interpret, evaluate, use and influence AI-generated intelligence.

This perspective is developed through the research and writing of Fotios Spyridonis, Senior Lecturer (Associate Professor) in Computer Science at Brunel University of London.

Why Accessible Intelligence?

Why Accessible Intelligence

AI can remove longstanding accessibility barriers while introducing new ones. An AI system may provide an accessible interface but produce explanations that are difficult to interpret. An intelligent assistant may reduce physical interaction demands while introducing new challenges for supervision and intervention. A recommendation may be clearly presented without giving someone sufficient information or opportunity to question it. Accessible interfaces do not necessarily make AI-generated intelligence accessible. Research in accessible computing, human-centred AI, explainable AI and responsible AI already addresses important aspects of these challenges. Accessible Intelligence connects these contributions by considering how accessibility affects meaningful participation throughout AI-mediated interaction. Its particular emphasis is on whether accessible interaction, meaningful understanding and verification, and exercisable human agency work together in practice. It builds on established scholarship and standards rather than proposing a new research field or framework.


♿ Accessible Interaction

Accessing and operating AI systems through suitable interfaces, modalities and assistive technologies.

📄 Meaningful Information

Ensuring generated descriptions, summaries and recommendations are relevant and useful in context.

🧠 Understanding & Evaluation

Enabling people to interpret AI outputs, recognise limitations and check consequential information.

👤 Exercisable Human Agency

Supporting accessible correction, intervention, alternatives and contestability in practice.


From individual capabilities to meaningful participation

An AI system may have an accessible interface, provide understandable explanations and offer mechanisms for correction or appeal. Yet these capabilities may not work effectively together. For example, a blind user may access an AI-generated image description but lack an accessible way to verify an important detail. Similarly, a person may understand an AI recommendation without having a practical opportunity to challenge it. Accessible Intelligence foregrounds this relationship: not simply whether individual accessibility, explanation and agency mechanisms exist, but whether they collectively support meaningful participation. This perspective builds on existing research and emerging standards on accessible and equitable AI systems.


Accessible Intelligence across computing

Generative AI

Accessible and contextually useful AI-generated content, with opportunities to identify and correct omissions.

Explainable AI

Explanations that support meaningful understanding and judgement for people with diverse abilities.

AI Agents

Accessible supervision, intervention and control as intelligent systems act increasingly autonomously.

Responsible AI

Ensuring safeguards such as transparency, oversight and contestability can be meaningfully accessed and exercised.


Publications

My research in digital accessibility, Human-AI interaction and inclusive interactive technologies informs the Accessible Intelligence perspective. Selected related publications are available on my Publications page.


Discuss and collaborate

Researchers

Interdisciplinary collaboration across accessibility, Human–AI interaction and responsible AI.

Students

Opportunities to explore accessible AI through PhD research, dissertations and student projects.

Industry

Collaboration on inclusive AI systems, accessibility evaluation and human-centred technology design.