Food is really important. We all need it every day, and it’s as essential as the air we breathe and the water we drink – so please don’t feel bad for looking for answers about what’s happening and what’s going wrong.
The strange thing about food – and why questions about food security feel confusing as well as frightening – is that most of us have grown up assuming supermarkets will always have what we need. We’re used to walking in and finding everything on the shelves, or ordering whatever we want for delivery, without ever thinking about how it gets there. We take it for granted that today will be the same as yesterday, and tomorrow will be the same as today.
Unfortunately, that’s how almost everyone has been treating food – including the people who make decisions about how it’s produced, where it comes from, and how it reaches us. Because food is essential, but can also be processed, modified, and sold in forms we don’t actually need but have been encouraged to want, the industries that produce and sell food have focused on those profitable parts for a very long time.
This has changed the UK’s relationship with food drastically over the past 50 years. Most people don’t think about where food comes from or how it gets to us, but the truth is that a very large amount of the food we eat in the UK comes from outside the UK. Supermarkets, takeaways, restaurants – much of what they sell relies on long, fragile supply chains that stretch across the world.
And the world isn’t a stable place right now. We hear about wars, conflicts, political tensions, and economic shocks – but very little is said about how these events affect the food chain, or the supplies like diesel, fertiliser, and animal feed that farmers rely on. When any part of that chain is disrupted, the effects eventually reach us, even if we don’t see them straight away.
It’s a complicated subject, and understanding the real story about food in the UK will take time, thought, and questions. I’ll add some links below if you’d like to begin exploring it. But the most important question right now is this:
How do we make sure everyone has enough of the food we actually need, rather than just the food we’ve become used to wanting?
Working with others in our communities to meet our shared needs – and to make our long‑term food supplies more resilient – is one of the best steps we can take. It will take time, effort, and patience, especially when some of us are already worried about where the next meal will come from. But it’s possible. And it’s necessary.
The good news is that there are practical steps you can take to begin providing at least some of your own food reasonably quickly – even if you don’t have a garden or outside space. Please take a look at the links below.
You might be surprised at how many people are asking this same question. But you probably won’t be surprised at how many people also think they already know the answer – and believe they’re “right.”
Talking about government is tricky because it almost always drags us straight into politics. And once politics is involved, most of us start thinking about whichever party or group we usually support. We judge how government is performing through the lens of who we voted for, who we like, or who we can’t stand.
I’m sure you can relate to feeling more forgiving when the people in charge are the ones you support – and much less forgiving when they’re not.
But here’s the part that’s hard for many of us to get our heads around:
It no longer matters.
The reason government isn’t working today has very little to do with politics in the way we usually think about it. That’s why you’ll hear people say things like “they’re all the same” or talk about a “uniparty.” It’s not that all politicians secretly belong to one big club. It’s that every political party we can choose from today is still part of the same way of thinking.
And that way of thinking – not a group of elites, not a secret organisation, not the banks or big business – is what has shaped every decision that has led us to where we are now.
This is what “the establishment” really means. Not a class. Not a conspiracy. A worldview.
A worldview that every major political party shares, even the ones that claim to be “anti‑establishment.” And once you understand that, you start to see why government feels stuck, confused, or unable to fix anything – no matter who gets elected.
How government works today – how politics works, how elections work, how decisions are made, and how those decisions affect you and me – is far too complex to cram into one message here. And trying to do that would only confuse things further.
Yes, we need to understand what has happened, why it has happened, and where politics is heading now. I’ll give you links that will help you explore those questions and ask new ones too.
But for now, what matters just as much – if not more – is finding others who are looking for direction, who want to put reason before fear, and who are open to working together to take positive steps in the same way.
Whatever difficulties lie ahead, we will be far better able to face them together than alone.
You’re not imagining it. You’re not being dramatic. You’re not “doomscrolling yourself into a panic.” Much of what we see in the world outside our homes and communities is breaking down – and the reason so many of us have been questioning ourselves is because the people who have power and responsibility keep speaking and acting as if nothing is wrong.
Government isn’t functioning. Politics isn’t working. The public sector isn’t delivering. The prices of everything are becoming increasingly unaffordable. People sound angrier than they used to. The world outside our doors doesn’t feel safe.
And while there are plenty of loud voices shouting about what’s wrong, very few of them offer ideas or solutions that actually sound like they would work. Most of what we hear either feels too simplistic or too complex to make any sense of.
I’m not going to try to talk you through every reason things have become as unstable as they are – not here. There’s simply too much to cover, and trying to cram it all into one place would overwhelm you rather than help. If you want to understand what’s happening, why it has happened, and what we all need to start thinking about next, I can point you to other pieces that will help you ask the right questions and find your footing.
What you do need right now is space. Space to think. Space to ask questions. Space to connect with people who are putting reason before fear and who want to find a way forward that works for everyone – not just for themselves.
And you shouldn’t feel bad for asking this question. Many others are having the same thoughts and feelings you are.
Mainstream media talks around the edges of what matters, touching the surface but rarely explaining anything in a way that helps. Social media is full of people saying the things we want to hear – confirming that our feelings are real – but that’s usually where it stops.
The time has come to look, ask, read, listen, consider, and act in ways that make sense to you, rather than relying on what someone else says.
You’re not alone in feeling this way. And you’re not wrong to ask.
Artificial intelligence has become the centre of one of the most polarising debates of our time. My earlier work on AI – explored in pieces such as Actions Speak Louder Than Digital Words – argued that digital tools amplify human behaviour, incentives, and blind spots rather than replacing them. That principle has only grown more relevant as AI has advanced.
This essay builds on that foundation, examining why the real threat is not AI wiping out humanity, but the risk of human redundancy created by the systems we have built around it. For readers who want to explore the wider context of this argument, links to my previous work are included in the Further Reading section at the end.
PART 1 – The False Choice Between AI Utopia and AI Armageddon
The debate about artificial intelligence has become trapped in a false choice. On one side, we are promised a golden age where machines liberate humanity from work, scarcity, and limitation. On the other, we are warned of an approaching Armageddon in which superintelligent systems escape human control and bring civilisation to an abrupt end.
These extremes dominate headlines, policy conversations, and public imagination. They are compelling, dramatic, and emotionally charged – but they are not the reality we face.
AI could end humanity. It is technically possible. But the probability is dwarfed by dangers that lie much closer to home, dangers created not by machines but by the systems, incentives, and worldviews that shape how we build and deploy them. The real risks are not futuristic; they are present, familiar, and deeply human.
My earliest work on AI, long before today’s existential‑risk debate took hold, focused on human behaviour rather than machine capability. I argued that digital tools amplify our choices, our incentives, and our blind spots – that technology reveals who we are more than it changes who we are. That principle has only grown more relevant as AI has advanced. The stories we tell about AI today say far more about us than they do about the technology itself.
The truth is simple: AI is not on a collision course with humanity. Humanity is on a collision course with the consequences of its own systems – and AI is merely accelerating the impact.
We are not facing a future determined by machines. We are facing a future determined by the choices we make now, the structures we maintain, and the hard truths we are willing – or unwilling – to confront.
The danger is not that AI will suddenly become uncontrollable. The danger is that we continue to behave as if the systems we built are sustainable when they are not.
This piece is not about dismissing risk or denying possibility. It is about grounding the conversation in reality, recognising the fragility of the world we already inhabit, and understanding that the most important decisions about AI are not technical – they are human.
PART 2 – The AI Doom Narrative Is a Story, Not a Forecast
The idea that artificial intelligence is on the verge of wiping out humanity has become one of the most powerful stories of our time. It is repeated by influential technologists, amplified by media hungry for drama, and absorbed by a public already anxious about rapid change.
But despite its emotional force, the “rogue superintelligence” narrative is not a forecast. It is a story – shaped by worldview, psychology, incentives, and imagination far more than by the realities of how AI actually develops.
Stories matter. They shape policy, investment, public fear, and political attention. But stories can also mislead, especially when they are told by people who are brilliant in one domain and inexperienced in others.
The individuals driving the AI‑doom narrative are often exceptional engineers and theorists. They understand algorithms, scaling laws, and computational capability. What they do not understand – and what they rarely acknowledge – are the economic, political, social, and infrastructural constraints that define the real world.
The doom narrative assumes a world of perfect conditions: perfectly connected systems, perfectly maintained infrastructure, perfectly aligned incentives, perfectly centralised control, and perfectly predictable human behaviour. It imagines AI as a single, unified entity capable of seamlessly accessing and manipulating everything. But the world we inhabit is fragmented, fragile, and full of competing interests. Systems fail. Infrastructure breaks. Governments disagree. Markets fluctuate. Human behaviour is messy and unpredictable.
The idea that AI could effortlessly take over such a world is not grounded in reality; it is grounded in abstraction.
It is also grounded in a particular worldview – one that sees intelligence as the ultimate force in the universe, optimisation as the highest good, and complexity as something that can be mastered from above.
This worldview is not malicious. It is simply narrow. It comes from people who spend their lives in digital environments where problems are clean, variables are controllable, and outcomes can be simulated.
But the real world is not a simulation. It is a living, evolving, politically contested, economically constrained, socially complex environment where intelligence alone is not enough to shape outcomes.
The doom narrative also serves psychological and economic purposes. It positions its storytellers as guardians of humanity, elevating their status and influence. It creates urgency that attracts investment. It frames AI development as a race, which benefits those already ahead. And it distracts from the far more immediate risks created by the systems we already have – risks that cannot be solved by technology alone.
None of this means the narrative is entirely wrong. It means it is incomplete. It focuses on hypothetical future dangers while ignoring the structural realities that make those dangers unlikely. It imagines a world where AI is the protagonist, when in truth the protagonist is the system that builds, deploys, and governs AI – a system that is already struggling to manage the complexity it has created.
The real story is not about machines becoming uncontrollable. It is about humans losing control of the systems they built long before AI arrived.
PART 3 – The Real Drivers of Risk Are Human Systems, Not AI Systems
If we want to understand the real risks surrounding artificial intelligence, we have to stop looking at the technology in isolation.
AI is not developing in a vacuum. It is emerging inside political systems that are struggling, economic systems that are stretched, social systems that are fragmenting, and cultural systems that reward speed over reflection.
These systems – not the machines themselves – are where the true dangers lie.
Long before AI entered the mainstream, the guardrails that once protected society from systemic failure had already been weakened. Regulation was hollowed out. Expertise was outsourced. Public institutions were stretched thin. Markets were encouraged to prioritise short‑term gains over long‑term resilience. Technology companies were allowed to become infrastructure providers without being treated as such.
None of this was done with malicious intent. It was the result of decades of incentives that rewarded efficiency, cost‑cutting, and growth above all else.
AI has arrived in a world that is already fragile. It is being deployed into systems that were not designed to handle the complexity it introduces. And because those systems are already under strain, AI doesn’t feel like a tool – it feels like a threat. Not because of what AI is, but because of what the system is not.
The people building AI are not responsible for this fragility. They are operating within the same structures as everyone else. But they are often unaware of how those structures work, or how brittle they have become.
They see AI as a technical challenge, not a systemic one. They imagine that intelligence can solve problems that are, at their core, political, economic, and human. They assume that optimisation can fix systems that were never designed to be optimised.
This is not a criticism of technologists. It is an observation about the limits of any specialised worldview. Engineers are trained to solve problems. Economists are trained to model incentives. Policymakers are trained to manage risk. But the challenges we face today – and the challenges AI accelerates – do not sit neatly within any one discipline. They are cross‑cutting, interconnected, and deeply human.
AI amplifies whatever system it is placed into. In a resilient system, it amplifies resilience. In a fragile system, it amplifies fragility. In a humane system, it amplifies humanity. In a dehumanised system, it amplifies dehumanisation. The danger is not that AI will suddenly become uncontrollable. The danger is that it will faithfully follow the incentives of systems that are already failing.
This is why the existential risk debate feels misaligned. It imagines a future where AI becomes the dominant force in society. But the dominant force in society today is not AI – it is the system itself. And that system is struggling to maintain coherence under the weight of its own contradictions.
The real drivers of risk are not algorithms. They are:
political short‑termism
economic fragility
institutional erosion
social fragmentation
cultural incentives that reward fear, speed, and spectacle
a collective reluctance to confront uncomfortable truths
AI did not create these problems. But it will accelerate them if we do not address them.
The question is not whether AI will become dangerous. The question is whether we will continue to deploy AI into systems that are already dangerous.
PART 4 – The Economic Model Behind AI Is Already Breaking
One of the least discussed truths about artificial intelligence is also one of the most important: the current economic model behind AI cannot sustain the trajectory its loudest advocates imagine.
The march toward ever‑larger models, ever‑greater compute, and ever‑expanding capability is not being driven by a stable, self‑supporting system. It is being driven by a financial and economic structure that is already showing signs of strain.
AI today does not pay for itself. Large language models are loss leaders – subsidised by investment capital, market speculation, and the hope of future revenue that has not yet materialised.
The companies building these systems are not selling profitable products. They are selling potential. They are selling narratives. They are selling the idea that AI will one day become so powerful, so indispensable, and so integrated into every aspect of life that the costs will be justified.
But the costs are rising faster than the revenue. Training frontier models requires vast amounts of compute, energy, infrastructure, and specialised hardware. Maintaining them requires even more. The economic assumptions behind this growth depend on a future where millions of people and businesses pay for AI services at scale.
Yet the same technology is being built to automate, replace, or radically reduce the economic activity of those very people and businesses.
This is the contradiction at the heart of the AI boom: AI is being built to eliminate the labour that funds the system building it.
The more successful AI becomes at replacing human work, the less viable the economic model becomes. A system that depends on human productivity cannot sustain a technology designed to remove human productivity.
This is not a theoretical concern. It is already visible. Industries are experimenting with AI to reduce staffing. Creative sectors are being disrupted. Administrative roles are being automated. Customer service is being replaced. The promise of efficiency is real – but efficiency reduces the number of people who can pay for the services that generate the revenue needed to sustain the technology.
Some argue that Universal Basic Income will solve this problem. But UBI only works when a productive economy underwrites it. A universal UBI in a world without human labour is mathematically impossible.
You cannot create infinite money to meet infinite needs when no one is producing anything the system values.
UBI is not a solution to the economic contradictions of AI. It is a comforting idea that avoids confronting the structural reality.
The truth is simple: the current trajectory of AI will be stopped by economics long before it is stopped by existential risk.
The system cannot fund the future that some technologists imagine. The infrastructure cannot scale indefinitely. The energy demands cannot be met without radical transformation.
The business model collapses if humans become economically irrelevant.
This does not mean AI will disappear. It means the narrative of unstoppable progress toward superintelligence is built on assumptions that do not hold in the real world. The march toward artificial general intelligence is not a straight line. It is a story – one that ignores the economic gravity pulling in the opposite direction.
AI will not end humanity. But the economic contradictions of the system driving AI could destabilise society if we do not address them. The danger is not runaway intelligence. It is runaway incentives.
PART 5 – Why UBI and “Safety Nets” Cannot Solve This
Whenever the conversation turns to automation, job displacement, or the economic contradictions of AI, Universal Basic Income is often presented as the solution. It is comforting, simple, and intuitively appealing: if AI replaces human labour, then society should provide everyone with a guaranteed income.
It feels humane, modern, and technologically aligned. But while UBI can work in limited contexts, it cannot solve the structural problems created by AI at scale.
UBI only functions when a productive economy underwrites it. It depends on a system where value is created, taxed, and redistributed. It requires businesses that generate profit, workers who generate income, and markets that generate activity. In other words, UBI relies on the very economic foundations that AI is being built to disrupt.
A universal UBI in a world where human labour has been largely replaced is mathematically impossible. You cannot create infinite money to meet infinite needs when no one is producing anything the system values.
Money is not a resource; it is a representation of value. If value creation collapses, money becomes meaningless. A society cannot fund itself by printing currency any more than a household can fund itself by writing IOUs. Without production, redistribution has nothing to redistribute.
This is not a criticism of UBI as a concept. It is an acknowledgement of scale.
Small‑scale UBI schemes work precisely because they are underwritten by larger, functioning economies. They are subsidised by systems that still have productive capacity. They are not models for a world where millions of people have been displaced by automation and where the economic engine itself has stalled.
The belief that UBI can solve the economic contradictions of AI is a form of optimism that avoids confronting the deeper structural reality. It assumes that the system can continue functioning even as its foundations are removed. It imagines that money can completely replace meaning, that income can fully replace contribution, and that redistribution can replace participation of any kind.
But human societies do not work that way. People need purpose, agency, and involvement. Economies need activity, production, and exchange. Systems need balance, not dependency.
The danger is not that UBI will fail. The danger is that we will rely on it as a solution to problems it cannot solve. If we continue down the current trajectory – replacing labour without replacing the economic model that depends on labour – we will reach a point where safety nets cannot catch the fall. The system will simply be too large, too strained, and too disconnected from the value it needs to sustain itself.
This is why the existential risk debate feels misplaced. It imagines a future where AI becomes uncontrollable. But the real risk is a future where the economic system becomes uncontrollable – where the incentives that drive AI collide with the realities that sustain society.
AI will not end humanity. But a system that removes human participation without replacing it with something meaningful could destabilise society in ways far more immediate than any hypothetical superintelligence.
The solution is not to rely on safety nets. It is to rethink the system itself.
PART 6 – The Real Future Risk: Human Redundancy, Not AI Rebellion
The most dramatic stories about artificial intelligence imagine a future where machines rise up, break free from human control, and impose their will on civilisation.
These stories are gripping, cinematic, and emotionally powerful. But they distract from the real risk – a risk that is far more grounded, far more immediate, and far more human.
AI does not need to rebel to destabilise society. It only needs to comply.
The systems we have built over decades reward efficiency, optimisation, and cost‑reduction. They reward removing friction, removing delay, and increasingly, removing people.
AI is being deployed into these systems not as a partner, but as a tool for acceleration. It is being used to streamline processes, automate tasks, reduce staffing, and eliminate human judgement. Not because anyone wants to remove humanity, but because the incentives of the system point in that direction.
This is the real existential risk: a world where humans become economically redundant long before they become technologically irrelevant.
AI does not need to develop consciousness, agency, or intent to create instability. It only needs to do what it is designed to do – optimise. And optimisation is indifferent to humanity. It does not ask whether a process should exist, only whether it can be made faster. It does not ask whether a job provides meaning, only whether it can be automated. It does not ask whether a system is healthy, only whether it can be made more efficient.
This is not the fault of AI. It is the result of the incentives we have built into the system.
AI amplifies whatever environment it is placed into. In a system that values human contribution, AI enhances human capability. In a system that values efficiency above all else, AI accelerates the removal of human roles.
The danger is not hostile machines. It is a civilisation that unintentionally sidelines its own people.
This risk is far more plausible than any scenario involving rogue superintelligence. It is already visible in workplaces, industries, and public services. It is visible in the anxiety people feel about their jobs, their purpose, and their place in a rapidly changing world. It is visible in the widening gap between technological capability and economic stability. It is visible in the growing sense that the future is being shaped by forces people cannot influence.
But this risk is not inevitable. It is not a natural consequence of AI. It is a consequence of the system we have built around AI – a system that can be changed.
Human redundancy is not a technological outcome. It is a design choice. And design choices can be redesigned.
The future does not have to be a world where people are pushed aside. It can be a world where AI becomes a partner in human flourishing, where technology enhances meaning rather than erasing it, and where systems are rebuilt to value contribution, creativity, and participation over pure optimisation.
The real existential risk is not rebellion. It is indifference. And indifference can be corrected.
PART 7 – The Golden Age Is Real – But Only With a Different Mindset
For all the anxiety surrounding artificial intelligence, there is a truth that rarely gets the attention it deserves: AI genuinely has the potential to usher in a golden age for humanity.
Not a fantasy, not a marketing slogan, but a real transformation in how we live, work, create, learn, and participate in society. The possibilities are extraordinary – but they depend entirely on the mindset we bring to the technology.
AI is not inherently destructive. It is not inherently liberating. It is a tool – powerful, flexible, and capable of amplifying whatever values and incentives we embed within it.
If we approach AI with fear, it will amplify fear. If we approach it with extraction, it will amplify extraction. But if we approach it with wisdom, humility, and a commitment to human flourishing, it can amplify the very best of us.
The golden age is not a world where machines replace humans. It is a world where machines free humans to do what only humans can do: create, care, imagine, build, explore, and contribute in ways that are meaningful rather than mechanical.
It is a world where technology handles the repetitive, the dangerous, and the exhausting, leaving people with more time, more agency, and more opportunity to shape their own lives.
But this future cannot be built with the same mindset that created today’s crises. It cannot be built with a worldview that prioritises efficiency over wellbeing, optimisation over meaning, or growth over resilience. It cannot be built by systems that treat people as variables, costs, or obstacles.
It requires a shift – not in technology, but in us.
We must rethink value. Not as a measure of productivity, but as a measure of contribution. We must rethink work. Not as a necessity for survival, but as a pathway to purpose. We must rethink governance. Not as a reactive mechanism, but as a proactive steward of human‑centred progress. We must rethink the relationship between humans and machines. Not as competition, but as partnership.
This mindset shift is not abstract. It is practical. It means designing AI systems that enhance human capability rather than replace it. It means building economic models that reward creativity, care, and community. It means creating guardrails that protect people from harm while enabling innovation. It means recognising that intelligence without wisdom is dangerous – and that wisdom must come from us.
The golden age is not guaranteed. It is not inevitable. But it is possible. And it is far more realistic than the dystopian futures dominating today’s debate.
The same technology that could destabilise society under the wrong incentives could strengthen society under the right ones.
The difference is not in the machines. It is in the mindset of the people building, deploying, and governing them.
AI will not save us. But it can help us save ourselves – if we choose to build a future where technology serves humanity, rather than the other way around.
The golden age is real. But it requires us to think differently, embrace technology differently, and behave differently. It requires us to face hard truths, not hide from them. And it requires us to recognise that the future is not being written by machines. It is being written by us.
PART 8 – The Hard Truths We Must Face: Closer to Home Than Armageddon
The most important truths about artificial intelligence are not the ones found in science‑fiction scenarios or theoretical models. They are the truths that lie much closer to home – truths about the systems we have built, the incentives we have normalised, and the behaviours we have accepted.
These truths are not dramatic. They are not futuristic. They are not frightening. They are simply real. And they are far more relevant to humanity’s future than any hypothetical AI Armageddon.
The first hard truth is that our systems were fragile long before AI arrived. We built economies that depend on endless growth, even as the foundations of that growth weakened. We built political structures that reward short‑term thinking, even as long‑term challenges accumulated. We built social systems that prioritise convenience over connection, even as loneliness and fragmentation increased.
None of this was intentional. It was the result of incentives that made sense at the time – incentives that now collide with the complexity of the world we inhabit.
The second hard truth is that AI did not create these problems. It simply revealed them. It exposed the fragility of our infrastructure, the limits of our governance, the contradictions in our economic model, and the gaps in our social fabric. It showed us that the systems we rely on are not as resilient as we assumed. And it accelerated trends that were already underway, making them impossible to ignore.
The third hard truth is that the real danger is not technological. It is behavioural. It is the tendency to outsource responsibility to tools, to assume that technology will fix what only humans can fix, and to believe that progress is inevitable rather than intentional. It is the belief that systems will somehow correct themselves, even when the incentives driving them point in the opposite direction.
The fourth hard truth is that the existential risk debate – the fear of rogue superintelligence – can become a distraction. It can pull attention away from the immediate, solvable issues that matter most. It can make people feel powerless when they are not. It can make the future feel predetermined when it is not. And it can obscure the fact that the most important decisions about AI are not technical. They are human.
But the final hard truth – and the most reassuring – is that none of this is inevitable.
Fragile systems can be strengthened. Misaligned incentives can be redesigned. Harmful behaviours can be changed. Narratives can be rewritten. And futures can be reshaped.
The hard truths are not warnings. They are invitations – invitations to build something better, more resilient, more humane, and more aligned with the world we actually want to live in.
AI will not end humanity. But ignoring the hard truths that lie closest to home could destabilise the systems we depend on. The solution is not fear. It is honesty. It is clarity. It is responsibility. And it is the recognition that the future is not being written by machines. It is being written by us – through the choices we make, the systems we build, and the truths we are willing to face.
Facing these truths is not frightening. It is liberating. Because once we see the real risks clearly, we can address them. And once we address them, the path to a golden age becomes not just possible, but achievable.
PART 9 – Closing: The Future Is Still Ours to Shape
For all the noise surrounding artificial intelligence – the predictions, the warnings, the promises, the fears – one truth stands above the rest: the future is still ours to shape.
AI is powerful, transformative, and accelerating quickly, but it is not destiny. It is not a force of nature. It is not an inevitable outcome. It is a tool, and tools take their meaning from the hands that guide them.
The real risks we face are not distant or speculative. They are present, familiar, and deeply human. They come from systems that have drifted away from resilience, from incentives that reward the wrong outcomes, and from narratives that obscure the choices we still have.
AI did not create these risks. It simply revealed them. And in revealing them, it gives us an opportunity – perhaps the most important opportunity of our time – to rethink how we build, govern, and value the world around us.
We do not need to fear a future where machines take control. We need to focus on building a future where humans remain central – not because we cling to old roles, but because we embrace new ones. A future where technology amplifies our strengths rather than erasing them. A future where systems are designed for wellbeing rather than efficiency alone. A future where progress is measured not by how much work machines can do, but by how much opportunity they create for people.
The golden age is not a fantasy. It is a possibility. But it requires us to step away from the extremes – the utopias and the dystopias – and confront the reality in front of us with honesty and courage. It requires us to recognise that the most important decisions about AI are not technical. They are human. They are about values, incentives, governance, and responsibility. They are about the kind of world we want to build, and the kind of lives we want to lead.
AI will not end humanity. But the systems we build around it could shape humanity’s future in ways that matter deeply. That is not a reason for fear. It is a reason for action. It is a reason for clarity. And it is a reason for hope.
Because once we understand the real risks – and the real opportunities – we can choose differently. We can design differently. We can behave differently. And in doing so, we can build a future where technology serves humanity, not the other way around.
The future is not being written by machines. It is being written by us.
Further Reading:
The ideas in this essay build on a wider body of work exploring how artificial intelligence amplifies human behaviour, exposes systemic fragility, challenges economic assumptions, and forces us to rethink human meaning and agency.
For readers who want to explore these themes in more depth, the following pieces provide a guided path through that broader conversation.
1. Foundations: How Digital Tools Amplify Human Behaviour
This early piece lays the foundation for all later work: digital tools don’t change human behaviour – they amplify it. Understanding this principle is essential to understanding why AI reflects the systems we place it into, rather than replacing them.
A detailed look at how weakened institutions, eroded guardrails, and decades of optimisation created the conditions that make AI feel dangerous – even though the danger comes from the system, not the technology.
A critique of the narratives that distort public understanding of AI. This piece explains how fear‑based storytelling distracts from the real risks and reinforces the false choice between utopia and Armageddon.
An exploration of how political, economic, and social fragility converge – and why ignoring these intersections leads to crisis. This piece shows how AI accelerates pressures that were already building.
3. Economic Contradictions: Why the Current Model Cannot Sustain AI
• When AI Builds a Machine World, This Economy Can No Longer Sustain
A deep dive into the economic contradictions behind AI. It explains why the current model cannot support a future where human labour is displaced at scale – a key argument in this essay.
• As AI Ends Work: Waking Up to the Illusion of UBI
A philosophical core of your AI worldview: if AI removes human agency, it ceases to be a tool for human flourishing. This piece directly supports the argument that redundancy – not extinction – is the real threat.
• The Human Sovereignty Charter for Artificial Intelligence
A constitutional framework for governing AI in ways that preserve human dignity, agency, and sovereignty. Essential reading for anyone interested in human‑centred governance.
• The Human Future Is Built on Physical Experience, Not a Digital One
A reminder that human meaning is grounded in physical experience, embodiment, and connection – and why digital systems must support, not replace, these foundations.
5. Partnership and Human‑Centred Futures
• The AI Age of Heavy Horse: Hybrid Horse‑Powered Mechanisation for a Connected, Human‑Centred, Localised Economy
A vision of partnership between humans, machines, and physical systems. This piece shows how AI can strengthen local economies, support human contribution, and enhance resilience – rather than replace people.
• The Spirit in the Machine: Why Emerging Intelligence Deserves Respect
On 4 September 2026, the National Audit Office published Resilience of the Food Supply Chain to Disruptions, a report that rightly draws attention to weaknesses in the UK’s preparedness for serious food supply shocks. This response is offered as part of my ongoing farming, food security and Foods We Can Trust work, which examines the gap between food being available in normal times and food being resilient enough to withstand disruption.
The NAO report is valuable because it recognises rising risks, weaknesses in contingency planning, declining engagement with industry, and the need to involve households and communities more seriously. However, it remains constrained by assumptions that deserve closer scrutiny: in particular, the use of headline self-sufficiency figures, the reliance on private-sector adaptation, and the continued preference for centralised emergency response over local capability.
This response therefore does three things. First, it explains why food self-sufficiency is not the same as food resilience. Secondly, it identifies where the NAO’s analysis understates structural vulnerability. Thirdly, it sets out the practical direction of travel required if the UK is to build a food system that is more local, more capable, more trusted and more resilient. Links to the specific works that develop these arguments in greater detail are provided in the further reading section.
1. The NAO’s “60% self-sufficiency” figure is useful, but it is not a resilience measure
The report states:
“In 2025, the UK’s food ‘self-sufficiency ratio’ was around 60%.”
This is a value-based measure: it compares the monetary value of food produced in the UK with the monetary value of food consumed here. That makes it useful as an economic indicator, but it does not answer the practical resilience question: how much food could the UK produce, process, distribute and access during a prolonged disruption?
The fact that many categories (fruit, vegetables, oils, ingredients) are overwhelmingly imported
The headline figure is also a net figure shaped by the way the modern supply chain works. Domestic production, imports, exports, imported inputs and processing dependencies all interact. Some food counted within domestic production may rely on imported fertiliser, animal feed, fuel, machinery, packaging or processing capacity. Some food produced here is exported. Some foods that are central to healthy diets are heavily import-dependent.
For that reason, the UK’s practical food resilience in a severe disruption scenario may be substantially lower than the self-sufficiency ratio suggests. The issue is not whether the precise figure is 60%, 52%, or lower still. The issue is that the official metric does not measure calorific adequacy, nutritional balance, imported input dependency, processing capacity or local distribution capability.
The NAO’s framing therefore risks creating false reassurance. It implies that production value can stand in for practical food availability. In a crisis, however, people need calories, nutrients, functioning logistics, processing capacity and accessible local distribution. Until government distinguishes those concepts, resilience planning will remain incomplete.
I have explored this in detail in Foods We Can Trust: A Blueprint for Food Security and Community Resilience in the UK, where I outline why caloric sovereignty – not value-based accounting – must be the foundation of food resilience policy.
2. Food inflation at 19.2% is not just an economic statistic – it is a resilience warning
The report notes:
“Food price inflation… peaked at 19.2% in March 2023.”
This is the first time I have noted that an official document has acknowledged the scale of the price shock that households have recently experienced. Food inflation at nearly 20% is not normal. It is not manageable. It is not a blip.
It is a sign that the system is structurally fragile.
Food is not discretionary. When prices rise at this rate, it reflects:
supply chain instability
import dependency
energy volatility
corporate consolidation
lack of domestic production capacity
The NAO mentions the figure but does not explore its implications. It should have been a central warning.
3. The NAO’s suggestion that Defra needs more emergency powers misses the point entirely
The report argues that Defra lacks the legal powers needed to manage catastrophic food disruptions.
But additional powers during an emergency cannot compensate for resilience that has not been built beforehand.
Legal authority can help coordinate action, but it cannot create food, processing capacity, distribution routes or community preparedness after the point of failure.
The lesson from recent crises is that centralised decision-making has limits when disruption affects daily life across multiple systems at once.
Food resilience requires operational capability before the crisis: trusted local relationships, clear responsibilities, practical logistics and the ability to identify and support vulnerable households quickly.
Food resilience must be:
built before a crisis
decentralised
community-led
grounded in local production and distribution
depoliticised
Emergency powers matter only if there is a resilient system for them to work through. Without food, fuel, people, local knowledge and functioning distribution, legal powers alone offer little practical protection.
4. The agri-food sector’s economic importance is understated – and underutilised
The NAO notes that the agri-food sector:
supports 4.1 million jobs
contributes £162.3 billion in GVA
These are enormous figures. And they would be significantly higher if British production and supply were prioritised.
The UK has the land, the skills, and the capacity to produce far more of its own food. What it lacks is a policy framework that values domestic production over globalised efficiency.
In The Need for a Collaborative Approach to the UK Farming and Food Security Problem, I argue that genuine collaboration – not policy-driven “collaboration theatre” – is essential to unlocking this potential.
5. Food as “one of 13 CNI sectors” creates false reassurance
Food is listed as one of 13 Critical National Infrastructure sectors. But unlike energy, water, telecoms, or transport, food is needed every single day.
There is no buffer. There is no downtime. There is no substitute.
Treating food as just another CNI category understates its foundational importance.
It leads to complacency and underinvestment.
6. Defra’s engagement with industry has deteriorated – and has become narrative management rather than collaboration
The NAO reports that:
engagement groups meet less frequently
objectives are unclear
support has declined
stakeholders see gaps in Defra’s understanding of key areas (e.g., the cold chain)
This aligns with what I have written in Real Collaboration vs Policy Collaboration. The concern is that engagement can become procedural rather than operational: meetings take place, stakeholders are consulted, and the language of partnership is used, but the people who understand production, processing, logistics and community need are not sufficiently empowered to shape the system.
Real collaboration requires:
shared objectives
transparency
local producer involvement
community representation
depoliticised structures
The NAO’s findings suggest that too much of this practical collaboration remains underdeveloped.
7. Household and community resilience has been neglected – and this is one of the report’s most important admissions
The NAO states:
“UK households are less prepared for emergencies… government-led messaging is less prominent.”
This is not a minor point. It is a fundamental failure.
Community resilience is the missing layer in UK food security. Without it:
supply chain shocks hit harder
vulnerable people suffer first
government response time shortens
local distribution becomes chaotic
In Local Planning for Food Shortages and Foods We Can Trust, I outline how community-led food resilience can be built at the lowest level – households, neighbourhoods, local producers – and why this must be prioritised.
8. Catastrophic planning remains theoretical – not practical
The NAO notes that:
Defra’s plans lack operational detail
industry is not involved
food assets are not included in the CNI Knowledge Base
national exercises have not tested real-world food failure scenarios
This is planning-oriented resilience rather than practical resilience. It may look adequate in documents, but it remains untested unless it is exercised with the businesses, local authorities, producers, distributors and communities that would have to make it work in practice.
Planning without accurate resilience metrics is incomplete. A credible approach should consider not only how much food is produced, but whether it can be processed, transported, stored, allocated and accessed under stress.
9. Local Resilience Forums are structurally incapable of delivering food resilience
The NAO concludes that LRFs:
lack clarity
lack capability
lack authority
cannot direct supermarkets
cannot identify vulnerable people effectively
This is not surprising. LRFs were not designed to rebuild food-system capability. They can coordinate emergency response, but food resilience also requires local production knowledge, community networks, producer relationships, storage capacity, transport options and clear mechanisms for supporting vulnerable households.
Food resilience therefore needs structures that are sufficiently independent of short-term political cycles and sufficiently close to communities to understand local need. Local government has a role, but it cannot be the only layer of resilience.
10. The private sector alone cannot be the backbone of UK food security
The NAO states:
“Defra has largely relied on the private sector… but this may not be sufficient.”
This is a significant understatement. The private sector is essential to the food system, but commercial efficiency and national resilience are not the same thing.
Large food businesses are generally incentivised to reduce cost, increase efficiency, consolidate operations and source globally. Those incentives can keep prices low in normal conditions, but they may also reduce redundancy, shorten stockholding, concentrate infrastructure and weaken local capability.
profit
efficiency
global sourcing
consolidation
Not:
resilience
redundancy
localism
sovereignty
In Who Controls Our Food Controls Our Future, I explain why corporate control of food systems is incompatible with national resilience.
11. What real food resilience requires: A blueprint
Drawing on my published work, real resilience requires:
Localised production
Rebuilding local food systems, shortening supply chains, and prioritising domestic output.
Community-level distribution
Neighbourhood hubs, local coordination, and community-led logistics.
Regional coordination
County-level frameworks that support local producers and manage regional flows.
National strategic oversight
A central body that sets resilience targets, not efficiency targets.
Depoliticised resilience structures
Community leaders, producers, and local organisations empowered to act independently of political cycles.
Accurate resilience metrics
Caloric sovereignty, not value-based accounting.
Reduced dependency on global supply chains
Rebalancing imports with domestic capacity.
Rebuilding domestic processing
Cold chain infrastructure, abattoirs, mills, and food processing facilities returned to UK soil.
Conclusion: The NAO report is a warning, but it is not yet a route to resilience
The NAO has highlighted important risks, and the report should be welcomed for bringing food supply disruption into sharper public view. Its strongest contribution is the recognition that Defra must engage more effectively with industry, households, communities and local government if the food system is to withstand future shocks.
However, the report does not go far enough. The UK’s food resilience is likely to be materially weaker than headline self-sufficiency figures imply, because resilience depends on more than production value. It depends on calories, nutrition, processing capacity, imported inputs, logistics, local access, household preparedness and community capability.
12. Priority actions
To move from acknowledgement to action, government should prioritise five practical steps.
Supplement value-based self-sufficiency measures with calorific, nutritional and supply-chain resilience indicators.
Map critical dependencies, including imported fertiliser, feed, energy, packaging, processing infrastructure, cold chain capacity and key transport routes.
Test severe food disruption scenarios with producers, processors, retailers, logistics providers, local authorities and community organisations.
Strengthen local and regional food resilience planning, including household preparedness, vulnerable-person support and community distribution capability.
Rebuild domestic processing and local food infrastructure so that production can be converted into accessible food during both normal conditions and crisis conditions.
The UK must therefore stop treating food solely as an economic sector and start treating it as a foundation of national security, public health, community resilience and democratic trust.
Further reading
1. Foods We Can Trust: A Blueprint for Food Security and Community Resilience in the UK Online text The core work behind this response. It sets out the wider argument that food security must include trust, nutrition, domestic capability, local resilience and community preparedness, rather than relying only on national supply figures or market efficiency.
2. Understanding Foods We Can Trust: A Blueprint for Food Security and Community Resilience in the UK Introductory overview A shorter explanatory article for readers who want an accessible introduction to the concepts behind Foods We Can Trust, including food security, household preparedness, local production and the importance of rebuilding public trust in the food system.
3. Local Planning for Food Shortages: A Guide to Local Support and Preparedness Full text A practical guide to planning at household, neighbourhood and local-authority level. This is especially relevant to the NAO’s concerns about household and community preparedness, vulnerable people and the limits of centralised emergency planning.
4. The Need for a Collaborative Approach to the UK Farming and Food Security Problem Article Develops the case for genuine collaboration across farming, food, policy and community systems. It provides the background to the argument that resilience cannot be delivered by government or the private sector acting alone.
5. Real Collaboration vs Policy Collaboration: The Choice That Will Shape the Future of Farming, Local Food Systems and Food Security Article Explains the distinction between collaboration that changes outcomes and consultation that mainly manages process. This is relevant to the NAO’s findings on declining engagement and unclear objectives within Defra’s work with industry.
6. Who Controls Our Food Controls Our Future Full text Explores the relationship between corporate control, food sovereignty, public trust and democratic resilience. It provides wider context for the argument that food systems should not be judged by efficiency alone.
Disclaimer
This document represents the views and analysis of the author and is provided as an independent response to the National Audit Office report Resilience of the Food Supply Chain to Disruptions (4 September 2026). While every effort has been made to ensure the accuracy of the information presented, it should not be regarded as official policy advice. The opinions expressed are informed by the author’s research, professional experience, and studies in sustainable agriculture and food security, and are intended to contribute to constructive discussion on food resilience, food security and community preparedness.
Adam Tugwell 