'Start small, think big': Scottish university lecturer Carl Dalby reveals how Britain can win the AI race

As Head of AI/Digital for a UK government agency, he bridges the worlds of public service and advanced technology. Over decades of technological leadership, Carl has held strategic roles in government, innovation agencies and digital programmes - combining governance, risk and transformation. As an artificial intelligence speaker, he is regularly called upon to translate complex AI topics into accessible, action-oriented insights for executive audiences.

In this exclusive interview with The AI Speakers Agency, Carl discusses the real challenges of implementing AI across organisations, the ethical frameworks that must accompany innovation, and how leaders can stay ahead in an era of accelerating change.

Carl Dalby serves as a guest lecturer at the University of Aberdeen in Scotland | The AI Speakers Agency

Q: Many UK businesses are now exploring AI technologies. What advice would you give to help them implement these tools successfully?

Carl Dalby: “Now, the concept of AI is still being defined and, in my opinion, will be defined for many decades to come. So, AI needs to be treated at this stage as more of a discovery — what it is and what it could be capable of doing within an organisation.

“But, as with any new discovery, start small. Pick a specific use case, whether it’s a function or a process or somewhere where there is availability of good amounts of data, and really focus on that use case and build it out.

“That’s the fundamental and critical piece when you’re looking at a major number-one tip for beginning the journey. Also, accept the fact that AI has been running across many organisations, whether we like it

or not, for many decades — it is not new. It has been around a very long time, whether you define that as machine learning or other deep learning capabilities around data.

“So, AI isn’t necessarily the new kid on the block, but it is something that has captured the imagination of many within organisations, employees and groups, and consequently it’s caught the imagination of leadership. What you want to do as a little tip is reaffirm that understanding of AI and what AI can and can’t do across the leadership teams, and get buy-in — that’s a really important part too.

“From my perspective, having a background in risk management as well, look at the risks of using AI. Evaluate them carefully across your security, AI data bias, AI errors, legal compliance, and all that good stuff, and just work out how you’re going to manage those in the context of your own organisation — that’s really important. Integration of AI with tools and processes is also a major challenge, but well worth looking at.

“As I say, start small and focus on specific use cases. Start with a proof of concept or a pilot — whatever your definition might be. My experience is you start with that proof of concept based on a challenging part of any business or process, and that really will pay a lot of dividends. If the proof of concept’s got a challenge statement that is common across many functions, it’s a good thing to reuse as a sales tool, shall we say, to get the AI message across.

“From a talent perspective, it’s a challenge. Start to look at the people internally who have passion and interest. Put skills on one side for a minute — are they evangelising AI techniques today and deeply interested in where the journey could go with it? Are they happy to challenge some of the misconceptions around AI in the public domain? That’s a really important bit — to get the right talent.

“And lastly, it’s never-ending. As with all good software and technology, it continually needs to be iterated. It feels like in the public domain right now, many of the major AI players are releasing phenomenal leaps in both quantity and quality of AI techniques almost daily. One of the biggest jobs we’ve got right now is keeping up with what’s actually happening.

“So, the final tip is all about trends — do the very best you can, not as a single process but by getting multiple people to share their experiences around what they see as being the next trends for your organisation in the use of AI. That’s a really powerful thing to do as well. Start small, get buy-in, and set realistic expectations on what AI can and can’t do — that’s the real focal point.”

Q: Ethical AI is becoming an important topic across the UK tech industry. What ethical challenges should organisations prepare for?

Carl Dalby: “Big question. My personal opinion — and it is a personal opinion — is that it’s a really interesting challenge. I find it fascinating that we are today talking about ethical challenges around the use of AI, and it’s a hot topic. I’ll come on to some of the areas that people need to be concerned about, look at, review, and consider.

“It’s interesting that we currently challenge ourselves on the ethical use of AI, but the world never stopped and took a breath to consider, for example, the ethical challenges of social media.

“It’s fascinating that we are now creating and running AI ethics committees and groups of that community and of interest — it’s absolutely critical in terms of governance.

“But I often reflect that we never stopped to look at the ethical challenges and concerns around the use of social media — and I include platforms like LinkedIn in that conversation.

“It’s a fascinating place to be right now in terms of the AI journey. When we look back at the impact social media has had across business, professionals, and the public worldwide — everything from democracies to data — it’s an interesting comparison.

“To answer the question directly, the areas in terms of ethical use to consider include bias within data and AI systems. AI has the ability to bring that bias into the forefront of people’s minds, and that’s a challenge that needs addressing.

“I’ve always been a complete advocate that AI must be explainable in a simple manner to all end users — whoever they may be. If you can’t explain the algorithm or what it’s trying to do, that’s a problem.

“Accountability is another major factor. When you ask for an output from any AI process — large language models being one flavour of AI among many — who’s accountable for that output? Who’s responsible when an AI makes a mistake? That needs to be built into every conversation.

“Data privacy is also crucial. There are some interesting questions around the data used to train AI models, and those questions should be asked early on. Security is another — AI systems are still vulnerable to cyber attacks and will be for many years to come, so robust security and ethical use of security must be prioritised.

“Legal compliance is a moving target. UK law is being written and developed as we speak. I’ve sat for nearly seven years on the government group for AI, and even now we’re still creating advisory positions on AI law in the UK. The EU’s AI Act provides a good template, and with new AI standards and safety institutes emerging in the UK and US, it’s clear that legal frameworks will shape much of how AI evolves.

“Lastly, there’s the question of unrealistic expectations. I spend a lot of time educating colleagues on what AI cannot do. Trust in AI is a dilemma right now — it’s human nature to trust what looks credible, but AI can produce incorrect or biased results. Managing those expectations ethically is vital.”

Q: The digital landscape is evolving fast. How can UK organisations stay ahead of the curve?

Carl Dalby: “I’d give a couple of pointers from my experience. Firstly, organisations need to really encourage and support a culture of innovation and true experimentation. Create an environment where business users can build proofs of concept — not just the tech platforms to do so, but easy access to play in a sandbox environment.

“Encourage continual learning. Provide ongoing training and access for employees at every level to gain new digital skills — from graduates to long-serving staff. Do it in a way that makes them comfortable to admit knowledge gaps and eager to learn.

“Work out which tech trends are relevant to your organisation’s focus and mission. Be aware that trends shift quickly — what’s popular today might fade tomorrow. The hype cycles of “metaverse” and “NFTs” are good examples. AI is different — it’s not a passing trend but an enduring transformation that will shape organisational DNA for decades.

“Run proofs of concept to test and learn at low cost. Partner up — engage both large and small tech firms, and collaborate with academia, which often leads the way on AI and machine learning innovation. Build those academic partnerships early.

“Finally, share the story — the good, the bad, and the ugly — across your organisation. Encourage open conversations, not just with supporters but with sceptics. Diverse viewpoints help avoid groupthink and create stronger, more balanced innovation strategies.”

Q: When you speak to audiences about AI, what do you want them to take away from your talks?

Carl Dalby: “I hope they feel the passion that this truly is a pivotal time for technology and AI. If I can encourage people to take a breath, step back, and understand that this is one of the defining moments in our digital evolution, then I’ve done my job.

“In my career, I’ve witnessed milestone moments — from the arrival of the desktop PC and the rise of the internet to the smartphone revolution and now AI. Each brought faster, smarter, more connected ways of working. But AI feels different — it’s the next seismic shift that will shape business, society, and daily life.

“I want audiences to pause, think, and recognise the opportunity AI brings — to exploit data responsibly, to enhance quality of life, and to build a better, more efficient world. It’s an exciting time to be part of this journey, and I want people to leave with the belief that this is the place to be — and the time to do it.”

This exclusive interview with Carl Dalby was conducted by Mark Matthews of The Motivational Speakers Agency.