After more than two decades of watching technology transform the world, I am less concerned about the machines than I am about the people designing, deploying and governing them.
There is a particular giddy kind of excitement that comes from watching technology evolve. It’s deeply personal and I’ve been fortunate to be part of the computer and digitalisation journey since the late 1980s and 1990s. Back then, I was a super geek! I was happy to spend hours understanding how systems worked, breaking things, rebuilding them and figuring out how to make them more resilient.
From floppy disks and IBM mainframes to data centres, virtual servers and cloud computing, I've had the privilege of experiencing technology evolve from the ground up.
I've broken Sun Solaris servers, built Linux networks sometime purely driven by inquisitiveness. I worked with Cisco infrastructure, secured TCP/IP and SQL servers, and worked on data centre failovers. I've seen technology change, adapt and transform the way we work and live.
The conversation swings between extraordinary optimism and predictions of catastrophe. Some see a technology capable of transforming human progress; others warn that it may ultimately threaten our existence.
There are legitimate concerns in both the opportunities and the risks. But I believe we are overlooking a more fundamental question.
What happens when the technology advances faster than our ability to take responsibility for how it is used?
That is not a question about machines alone. It is a question about us.
We have been here before
Artificial intelligence did not arrive with ChatGPT, Gemini Ai or Claude Ai, nor did the ambition to build intelligent machines begin in Silicon Valley.
Alan Turing's work in the middle of the twentieth century helped establish some of the intellectual foundations for modern thinking about machine intelligence. In 1956, the Dartmouth workshop helped establish artificial intelligence as a formal field of research. Since then, progress has come in waves, each building on the work of researchers and engineers before it.
Logic Theorist in 1956 to IBM Watson winning Jeopardy! in 2011, the breakthroughs in deep learning, AlphaGo and the introduction of Transformer architecture in 2017, we have seen extraordinary progress.
Every one of these developments has involved people. Researchers, engineers, businesses, investors and governments have all contributed to the journey.
So, when I hear people saying that AI will be the destruction of mankind, as though it is an independent force moving through history, rather than a technology shaped by human choices.
I'm not dismissing the risks. We absolutely need to understand them. But we also need to be honest about where responsibility sits.
If we want to prevent harm, we need to look closely at human decisions, not just the machines. The dates matter, but so does what lie behind them.
Every breakthrough depended on human decisions what to research, where to invest, what to build and how to deploy it.
If an AI system causes harm, we need to understand the technology. But we must also examine the data, the design, the incentives, the testing, the deployment and the decisions of those responsible for it.
The machine may be new. The question of responsibility is not.
The data beneath the promise
Let's talk about data, because it matters
One area I believe deserves much more attention is AI data.
We can build the most sophisticated AI model in the world, but if the data behind it is inaccurate, poorly managed, biased or exposed, we create problems before the system even has a chance to deliver value.
Having worked with infrastructure, databases, networks and cybersecurity for many years, I've learned that data is never just a technical issue.
It is about trust.
Where did the data come from? Who owns it? Do we have the right to use it? Is it accurate? How is it protected? And what happens when it is shared with third parties or fed into AI systems?
These are questions every organisation needs to ask.
Think about the information people share with AI tools every day. Personal details, business information, customer records, confidential documents and sometimes information that should never have been uploaded in the first place.
How many people genuinely understand where that data goes, how it might be processed or what controls are in place?
This is where education and accountability become so important.
We need to make AI easier to understand, not more intimidating. People should feel confident asking questions before they use a tool, not discover the risks after something has gone wrong.
And for businesses, AI governance cannot be an afterthought. Data protection, access controls, monitoring, testing and clear accountability need to be considered from the beginning.
My early career was rooted in infrastructure and systems. I learned how information moves across networks, how databases support operations and how failures in one part of an environment can affect everything around it. Over time, I came to understand that protecting technology also means understanding the information it holds and the decisions it enables.
AI makes that responsibility more complicated.
An organisation may introduce an AI system without fully understanding the origins of its training or operational data, the permissions attached to it, the information it retains or the third parties through which that information travels.
The questions are not particularly glamorous, but they are essential.
Where did the data come from? Was it collected and used appropriately? Is it accurate and representative? Does it contain personal or confidential information? Who can access it? How do we challenge an incorrect output? And who is accountable when a decision made with the assistance of AI causes harm?
These are not questions that can be answered by the technology department alone.
They belong in the boardroom, alongside legal, risk, security, data and operational leadership.
The danger is not merely that an AI model might make a mistake. It is that an organisation might place its trust in a system without understanding the conditions under which that system can fail.
Trust in AI cannot be separated from trust in the data, the processes and the people behind it.
Fear is no substitute for evidence
There is a temptation, whenever a powerful new technology emerges, to reach for the most alarming possible outcome.
We have seen this pattern before. The Y2K problem prompted serious concern about the ability of computer systems to handle the transition to the year 2000. Some of the risks were exaggerated and organisations invested substantial effort in identifying and addressing them.
The lesson I draw is not that warnings should be dismissed. It is that risks deserve to be examined, tested and addressed with discipline.
We should approach AI in the same spirit.
There will be genuine risks; misuse, fraud, privacy breaches, insecure systems, discriminatory outcomes and failures in critical processes. There may be other consequences that we have not yet fully understood.
But we do not improve our ability to manage those risks by relying on speculation or dramatic predictions unsupported by evidence.
We need research, transparency, independent testing and a willingness to acknowledge uncertainty.
We also need to resist the idea that regulation alone will resolve the problem.
Governments have an important role in establishing safeguards and protecting the public. We all know that legislation takes time, and technology continues to develop while policymakers deliberate.
We need to setup clear accountability for the leaders/CEOs of the AI companies; they need to take security and controls seriously.
Organisations cannot simply wait for a complete set of rules before deciding how they will use AI responsibly.
They need to make those decisions now.
Don't overlook the people who already have the experience
Something else has been on my mind.
There are so many talented people in cybersecurity, technology and risk who have been made redundant or are being underutilised, despite having years of practical experience that underpin today's digital economy.
Why aren't we making better use of that knowledge?
I've always believed that you learn from the past to build the future. We shouldn't dismiss people simply because their experience comes from an earlier generation of technology.
I have worked across industries and countries where the risks associated with technology are shaped by far more than software. Infrastructure, human behaviour, cultural differences, geopolitical circumstances and business pressures all influence how a system is designed and how it can fail.
Someone who has worked with legacy infrastructure may see a dependency that a new implementation overlooks. An individual that has managed an incident understands the difference between a plan on paper and a decision made under pressure. An experienced person that has worked across jurisdictions may recognise that a control that works in one environment cannot simply be assumed to work in another.
This is not an argument against new expertise. It is an argument for bringing different forms of expertise together.
The future should not be built by choosing between experience and innovation. It should be built by connecting them.
We should be using experienced professionals to advise boards, support governments, shape technology programmes and mentor those entering the industry.
There is a great deal of knowledge in the workforce that could be put to better use.
Education should come before the crisis
AI literacy cannot be treated as a privilege for those who work in technology.
Let's make AI education accessible to everyone.
AI education shouldn't be limited to technical specialists or large organisations with big budgets.
The general public needs practical guidance, too.
How do you recognise an AI-generated scam? What information should you avoid sharing? How do you check whether an AI answer is accurate? When should you question a decision made with AI?
These are not abstract questions. They are becoming part of everyday life.
I want people to feel confident using technology, not frightened by it.
That includes children, families, small businesses and people who may not have had the same opportunities to access digital education.
We need to teach people how to use AI responsibly, how to recognise risks and how to think critically about the information they receive.
We need to do it in a way that is simple, practical and accessible.
A message to boards, governments and technology leaders
Don't wait for the next major incident before asking the difficult questions.
Don't rely only on polished presentations or assurances that everything is under control.
Bring in people who have practical, hands on experience. People who understand how systems work, how risks emerge and what happens when technology doesn't behave as expected.
Please don't overlook those who have spent their careers building the foundations on which today's innovation depends.
Their experience matters.
The next generation needs their mentorship. Organisations need their knowledge.
I'm still as excited about technology as I was when I first started working with computers. I see so much potential for responsible use of AI. Innovation creates opportunities, improve lives and help businesses solve real problems.
But excitement needs to come with responsibility.
We need to build AI with clear boundaries, appropriate data governance and human accountability.
We need to educate rather than intimidate.
Most importantly, we need to remember that the future of AI is not simply about what machines can do.
It is about the decisions we, as humans, make today.
That's where the real work begins.