I love it when my worlds of passion collide.
Recently, I’ve been involved in some fascinating conversations about AI agents, ethics, and how organisations should identify and manage bias within AI systems. One theme kept emerging: what framework should organisations use to define fairness?
For many organisations in the UK, the answer is straightforward. The Equality Act provides an established legal framework for protected characteristics and discrimination. It’s a logical place to begin when designing, procuring, or governing AI systems.
But those conversations highlighted something I’ve been passionate about for many years.
The people our frameworks don’t see
If organisations use the Equality Act as the benchmark for identifying and mitigating bias, what happens to groups that aren’t recognised within it?
Care-experienced people are one of those groups.
We know that people who have grown up in care continue to face disproportionately poor outcomes across education, employment, housing, physical and mental health, and interactions with the criminal justice system. These aren’t isolated statistics they represent systemic disadvantage that can follow people throughout their lives.
Yet, despite growing recognition of these challenges, care experience is not a protected characteristic under the Equality Act.
AI will only ever be as fair as the frameworks behind it
One of the biggest misconceptions about AI ethics is that bias is purely a technical problem. It isn’t.
AI systems learn from data, but they are designed, governed, and assessed by people. The frameworks we choose determine what we measure, what we test, and ultimately, who we protect.
If organisations only assess bias against legally protected characteristics, there is a real risk that care-experienced people become invisible within AI governance.
This isn’t because organisations don’t care.
It’s because our frameworks influence our thinking. They shape procurement questions, impact assessments, governance processes, and assurance activities. If a group isn’t explicitly included, it is far less likely to be considered consistently.
As AI becomes increasingly involved in recruitment, education, housing allocation, financial services, healthcare, and public services, this becomes more than a legal discussion…it becomes a question of social justice.
A stronger case than ever
For years, campaigners have argued that care experience should become a protected characteristic because of the persistent inequalities faced by this community.
The rise of AI adds another compelling dimension to that argument.
If the Equality Act is increasingly becoming the foundation for ethical AI in the UK, then recognising care experience as a protected characteristic isn’t simply about strengthening legal protections. It’s also about ensuring the next generation of technology doesn’t unintentionally reinforce the inequalities that already exist.
Fair AI requires fair frameworks.
And fair frameworks should reflect today’s understanding of disadvantage….not yesterday’s.
Looking ahead
The conversation about AI ethics shouldn’t only be about algorithms, transparency, or explainability.
It should also challenge us to ask whether the legal and ethical frameworks we rely on truly represent everyone who needs protection.
Perhaps the rapid growth of AI gives us another reason to revisit the question:
Should recognising care experience as a protected characteristic become a priority, not only for improving equality today, but for ensuring fairness is built into the technologies that will shape tomorrow?
Because if we’re serious about creating responsible AI, we also need to be serious about who our definition of fairness includes.
Leave a Reply