GUEST COLUMN:
Tim Bashford
Research Lead, Wales Institute of Digital Information
University of Wales Trinity St David
AI is often talked about as though it is something the public sector is destined to trail behind on. I’m not convinced that is true. In fact, from what I see through my work at the Wales Institute of Digital Information, the picture is far more nuanced.
The divide is not public versus private. It is between organisations that are large, digitally mature and well resourced, and everyone else. And in that “everyone else” category, there are a great many private sector organisations, particularly smaller businesses, that are no more prepared for AI than the least mature parts of the public sector.
Our Institute is a partnership between the University of Wales Trinity St David, the University of South Wales and Digital Health and Care Wales. We work closely with the Welsh NHS on innovation, research and development, and training.
Many of the challenges facing healthcare today are innovation problems, and they cannot simply be scaled out of. They demand new approaches, new ways of thinking and often new technical capability. AI sits firmly within that space. We spend a lot of time not only looking at the technology itself, but also at how people and organisations can adopt it safely and confidently.
What is striking is that the common narrative – that public services are always far behind the private sector – doesn’t really hold. In Wales, public services actually have comparatively mature governance and support structures for AI. There are, of course, variations across organisations. Legacy systems and legacy capacity can slow adoption. But we see real pockets of strength too. Meanwhile, the private sector is not uniformly ahead. Yes, there are large organisations, particularly in areas like fintech and medtech, that are doing a very good job of adopting AI. But there is also a long tail of small and medium-sized enterprises that are struggling with readiness, skills and investment. The gap, therefore, is not one of public versus private, but of scale and maturity.
This is where I believe higher education can play a more active and practical role. Too often, universities are seen solely as the homes of the highfaluting algorithmic stuff – the deep technical research and complex computer science that most people don’t encounter day to day. While that research is important, it isn’t the only contribution we can make. In fact, I would argue that our role in training and skills development is just as critical, particularly as organisations across Wales start to think about how AI could be used responsibly.
We are already seeing the need for training that sits well beyond advanced algorithm design. Public sector staff need to understand how to use AI tools, how to question them, how to interpret their outputs and how to apply them in ways that are ethical, transparent and safe. At the same time, they need to avoid over-reliance. AI can be convincing even when it is wrong, and complacency is a risk. So is the risk of unequal outcomes where data carries ingrained bias, or where languages such as Welsh are not treated equally because the underlying models have not been trained with the right balance. These are real-world concerns that require real-world training.
There is also an emerging need for what we might call prompt literacy: the ability to use AI effectively within everyday work. For many people in office environments, tasks such as generating material, synthesising information or producing early drafts are already becoming part of their routine engagement with AI tools. Used properly, these tools can free up time and improve productivity. Leaving that capability untapped simply because there has not been enough training or confidence-building seems like a missed opportunity.
That is why we are starting to develop small, accessible CPD packages to help organisations build the skills they need. These are designed to be practical and usable, not theoretical. They focus on what employees actually need to know to make their jobs easier and to use AI responsibly. They also consider ethical issues such as fairness, transparency, the risks of misinformation and how to maintain trust, because trust, once lost, is very hard to regain.
In my view, this is where higher education can make a significant difference. We are well placed to support organisations that want to move beyond the initial hype and into thoughtful, well-governed adoption. We can bring research expertise, but we can also translate that into training, guidance and frameworks that work for people who are not specialists. And importantly, we can work with organisations of all sizes, not just those with large budgets or in-house digital teams.
AI will continue to evolve, and the skills needed to work with it will evolve too. The public sector in Wales is right to take a measured approach: start with the problem, understand the purpose and build the capability before rushing to deploy tools. If we can support that journey with training and practical guidance, rather than only with high-level technical research, then higher education can help ensure that organisations across Wales – not just the biggest ones – can use AI safely, confidently and for the right reasons.
Tim Bashford talks about this and more in the Government and Not for Profit podcast episode Ready for AI. Listen here.
