During bath time, my child happily announced: “Mum, I’m making so many bubbles for my bath so that my body will be clean.” 🤣😂
Without giving it much thought, I replied:
“If you make all the bubbles in the world without thoroughly scrubbing your body, your body will still remain dirty.”
That response came spontaneously to me, but later on, I found myself thinking deeply about it.
Bubbles can make a bath look exciting.
They can fill the tub, sparkle under the light, and create the appearance that something important is happening.
But bubbles alone do not complete the real work.
And suddenly, I realized how closely this relates to the way we are building artificial intelligence today.
We are surrounded by countless exciting AI announcements.
New tools are being launched daily.
Models are becoming faster daily.
Systems are generating images, writing documents, supporting decisions, automating work, and influencing how people access education, healthcare, employment, financial services, and public information.
There are indeed bubbles everywhere.
But beneath all that excitement, we must ask a more important question:
Are we thoroughly doing the work required to make AI fair, inclusive, accessible, and useful for everyone?
This question has never been more important.
Across the world, experts are warning that AI can unintentionally exclude the very people it is meant to help when inclusion is overlooked. Canada’s Human Rights Commission has cautioned that biased data and design can reproduce systemic discrimination and leave diverse communities behind. Researchers at Stanford University Graduate School of Business have found racial disparities in AI-powered hiring systems, while UN Women has highlighted evidence of gender and racial bias across evaluated AI models. Canada’s National AI Strategy emphasizes that AI should benefit all Canadians, and the United Nations continues to warn that unequal access to AI and limited language representation could deepen global inequalities.
These are reminders that innovation alone is not enough.
Because if we build intelligent systems without considering people from minority and marginalized communities, we may create impressive technology that still fails the very people it was supposed to serve.
A hiring system may process thousands of applications quickly, but what happens when qualified candidates from certain racial groups are unfairly screened out?
A voice assistant may respond within seconds, but what happens when it struggles to understand particular accents, speech patterns, or disabilities?
An AI health tool may analyze information efficiently, but what happens when the data used to train it does not adequately represent people with darker skin tones or underserved communities?
An educational assistant may personalize learning, but what happens when children without reliable internet access, appropriate devices, language support, or accessible interfaces cannot benefit from it?
The technology may appear advanced, but underneath the bubbles, the deeper work remains unfinished.
This is where UX design and human-centred AI become essential.
As designers, researchers, developers, and decision-makers, we must move beyond asking:
“Can we build it?”
We must also ask:
Who was included in the research?
Whose experiences shaped the training data?
Who might be misunderstood or misrepresented?
Can people with disabilities use the product?
Does it work across different languages, cultures, skin tones, accents, incomes, and locations?
Were affected communities invited to help design and test it?
Is there a human being available when the system gets something wrong?
Inclusive AI cannot be created by assuming we understand everyone. It must be built with people, not simply for people.
This is also the heart of my Chapter 21 contribution, “Empathy by Design: How AI and UX Shape a More Inclusive Digital World,” in AI Everywhere, Volume 1.
In the chapter, I explored why empathy, accessibility, representation, and inclusive design must be part of AI development from the beginning.
They should not be treated as optional additions after a product has already been built.
Because inclusion is not decoration.
It is not a final accessibility checklist.
It is not something we sprinkle over technology to make it appear responsible.
It is part of the real work.
My child thought more bubbles would automatically create a cleaner bath.
And in her sweet little statement, I found a powerful reminder for our rapidly changing world:
More AI does not automatically mean better AI.
Faster AI does not automatically mean fairer AI.
Smarter AI does not automatically mean more human-centred AI.
We can create all the impressive features in the world, but if we do not thoughtfully include the people our technology will affect, something important remains unclean beneath the surface.
The future of AI should not only be powerful.
It should be fair.
It should be accessible.
It should understand human differences.
Most importantly, it should leave no one behind.
Government of Canada Interdepartmental AI Community of Practitioners
References
AI Hiring Tools Can Yield Racial Bias and Systemic Rejection
Canada’s National Artificial Intelligence Strategy: AI for All
#BlessingSeries #InclusiveAI #HumanCenteredAI #ResponsibleAI #EthicalAI #ArtificialIntelligence #UXDesign #InclusiveDesign #Accessibility #DigitalInclusion #DesignThinking #UserExperience #UXResearch #EmpathyByDesign #AIForGood #AlgorithmicBias #TechForGood #ProductDesign #DiversityInTech #FutureOfAI
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.