The most powerful AI system is worthless if people cannot use it, trust it, or understand it. Human-centred AI design is the discipline of placing people — their needs, capabilities, and limitations — at the core of every design decision. It is an approach that recognises technology as a tool in service of human goals, not an end in itself. As AI systems become more capable and more pervasive, the principles of human-centred design become not just desirable but essential.
Core Principles
Human-centred AI rests on several interconnected principles. First, transparency: users should be able to understand, at an appropriate level, how an AI system reaches its conclusions and what data it relies upon. This does not mean exposing every algorithmic detail to every user, but it does mean providing meaningful explanations that allow people to make informed decisions about whether to trust and act on AI outputs.
Second, inclusivity. AI systems must be designed to work for the full diversity of people who will interact with them. This means actively testing for and mitigating biases related to race, gender, disability, age, and socioeconomic background. It means involving diverse voices in the design process itself — not just as test subjects, but as co-creators who shape the system's purpose, behaviour, and boundaries. Inclusivity is not a feature to be added; it is a standard to be maintained from the outset.
Designing for Trust
Trust is earned through consistent, predictable behaviour and honest communication about limitations. A human-centred AI system should be clear about what it can and cannot do, gracefully handle edge cases, and provide users with meaningful control over how they interact with it. When errors occur — and they will — the system should fail safely, provide clear explanations, and offer pathways for human intervention. Over-promising and under-delivering is the fastest way to destroy the trust that AI systems need to be effective.
The AI Board embeds human-centred design thinking throughout its qualifications and endorsement criteria. We believe that every professional working with AI — whether designing, deploying, or overseeing it — should understand these principles and be equipped to apply them in practice. By making human-centred design a core competency rather than a specialist niche, we can ensure that AI technology develops in ways that genuinely serve the people it is intended to help.
Back to Insights