There is a moment, when you read a letter that says what you have been circling for years, when the argument suddenly feels both obvious and urgent. That was my reaction to Aneet Morar’s response to the Guardian’s editorial on Russian disinformation. Yes, the warning from Andy Burnham about the Kremlin was important, and yes, it captured something real about foreign interference in British democracy. But as Morar argues, it really only tells half the story. The harder truth is that disinformation has become just one symptom of a much deeper failure: we have allowed a handful of private companies to become critical national infrastructure without ever treating them as such. They are the pipes and cables through which public conversation now flows, the town squares where politics, friendship and commerce meet, the libraries, the newspapers, the advertising hoardings, the church noticeboards of our age. Yet unlike any other form of critical infrastructure, they are run according to engagement metrics, not public duties, and their owners answer first to shareholders, not citizens. We have built the modern public square on private land, and then we are surprised when the rules of that square are written by those who profit from the loudest, most extreme and most emotionally charged voices. The Kremlin’s bot armies are genuinely dangerous, but they are only able to do so much damage because the infrastructure itself has been designed to reward amplification, not truth. If we keep focusing only on the foreign actors pressing the buttons, we will never ask the uncomfortable question of why the buttons work so well in the first place.
That question becomes more painful when you look at the full spectrum of harm caused by engagement-driven feeds. We know from research, from clinical practice and from the testimony of teenagers themselves that these systems are damaging young people’s mental health. The same algorithmic mechanics that surface a Russian smear campaign also surface images and videos that make a fourteen-year-old feel profoundly inadequate in her own body. The same recommendation engine that pushes a user from vaccine hesitancy to conspiracy theory can push a lonely young man from online resentment to far-right radicalisation. It is not a coincidence, and it is not a series of separate problems. It is one architecture producing one kind of outcome: content that keeps people on the platform, regardless of what that content does to them or to the wider society. When the Guardian’s analysis of the disorder in Belfast and Southampton showed how far-right violence was fuelled by online falsehoods and amplification, it was a reminder that the consequences are not limited to the inside of our heads. People have been harassed, hospitals have been put on alert, homes and businesses have been attacked, communities have been terrified. And all of this is the predictable result of systems that are built to reward outrage because outrage holds attention. We have spent years treating these outcomes as if they were exceptional moral failures of individual bad actors, rather than the ordinary operating conditions of a machine whose entire purpose is the exploitation of human attention. Until we stop marvelling at the symptoms and begin examining the machine, we will continue to be surprised by the very disasters that the machine is designed to produce.
The newest technologies only make this more urgent. The next generation of wearable devices, including Meta’s smart glasses, is normalising a world in which recording other people without their knowledge or consent is a routine part of daily life. Imagine standing in a supermarket queue, on a bus, at a school gate, at a gym, at a protest, at a hospital appointment, and never knowing whether you are being captured and uploaded. Imagine the chilling effect on ordinary behaviour, on political organising, on the simple freedom to be imperfect in public. The camera is no longer a visible, respectful, deliberate act; it is a function of the frames on your face. And once these devices are connected to the same engagement-driven platforms, every face in the background becomes data, every mistake becomes a potential moment of viral humiliation, every private conversation becomes potentially public. The harm is not only to victims of specific recordings. It is the slow erosion of trust itself, the feeling that we are always being watched by someone who might not even know they are watching us. This is not a science-fiction nightmare. It is already here, and it is being rolled out, as so much else has been rolled out, in a regulatory vacuum. We are allowing a handful of companies to decide, in effect, what consent means, what privacy means, what a public place means, with no democratic mandate and no independent oversight. And we are doing so at a speed that leaves us no time to think seriously about the kind of society we want to live in.
The government’s response to disinformation has been real, but it is not enough. Yes, there are units, there are strategies, there are occasional announcements about countering hostile state activity. And there is something in the online safety agenda that gestures toward a better world. But the fundamental problem remains: platforms that profit from spreading harmful content still do not face meaningful consequences. An enforceable duty of care, penalties that are proportionate to the scale of the damage they cause, and genuine transparency about how their algorithms decide what we see would not be a radical agenda. They would simply be the basic regulatory minimum for an industry that now holds enormous power over what millions of people believe, how they vote, how they feel about themselves, and how they treat one another. We are not asking for an end to the internet, or for censorship of unpopular opinion, or for the impossible task of making every speech safe. We are asking for a system in which someone is accountable. At the moment, when a platform amplifies a lie that leads to violence, there is no one who carries responsibility. The algorithm is a black box, the company says it is merely a conduit, the regulator wrings its hands, and the next disaster is already being planned. We would never accept that from a water company, from a bank, from an electricity grid, from a railway network. Why do we accept it from the platforms that carry the very social fabric itself? The answer is not that the problem is too hard. The answer is that the platforms have spent enormous resources making sure that no one ever tries.
I spent my career in the payments industry, and I know what responsible infrastructure looks like. In that world, scale is not an excuse for harm. If a payment system fails, or processes a fraudulent transaction, or is used for money laundering, there are rules. There is independent oversight. There is liability when things go wrong. The infrastructure is not allowed to grow so large that it becomes critical and then say, “We are too big and too complex to be regulated, and anyway we are just a pipe.” The opposite is true. The more vital and central the infrastructure, the more rigorous the requirements. That is what makes it safe enough to be trusted with the flow of money, the medical records, the identities, the livelihoods of ordinary people. The platforms that shape our beliefs hold something even more important than our money: they hold our attention, our memory, our sense of reality. And yet they are held to a lower standard than the companies that move our bank balances from one account to another. There is no good reason for this. It is not as if the technology is too new to understand. It is not as if the harms are impossible to measure. It is not as if no one can see what is happening. Everybody can see it. We have simply allowed this new industry to persuade us that its business model is a force of nature, that its algorithms are beyond human comprehension, that its power is somehow inevitable. It is not inevitable. It is the result of choices, choices made in boardrooms, choices made in government, choices made in our own acceptance of the status quo. If we can regulate the transmission of money, we can regulate the transmission of meaning. If we can demand that a bank be transparent about why it froze an account, we can demand that a platform be transparent about why it boosted a lie to millions of people.
Tackling the Kremlin’s bot armies matters. It matters for national security, for democracy, for the integrity of elections and for the safety of communities that are deliberately targeted by hostile actors. But we will never get to the root of the problem by chasing individual bots or taking down individual accounts. We need to treat the cause rather than the symptoms. And the cause is structural: we have allowed profit-driven amplification to become the organising principle of public discourse. Until we change that principle, the same harm will keep returning in different forms. A foreign government will use one set of tools; a domestic extremist will use another; a troubled teenager will find their own path into the abyss; every one of us will be shaped, in ways we cannot always see, by systems that have no allegiance to anything except their own growth. There is a better way. It begins with naming exactly what these platforms are. They are critical national infrastructure. They are as important to our public life as roads and water and electricity. And they should be held to the standard of critical national infrastructure: with duties of care, with independent oversight, with liability for harm, and with rules that put the safety of people ahead of the profits of a few. It is not a question of being anti-technology, or anti-business, or anti-free speech. It is a question of being pro-democracy, pro-community and pro-human. The scale of the task is enormous, but so is the cost of doing nothing. We have already seen what that cost looks like in the mental health crisis, in the radicalisation of young men, in the violence on our streets, in the erosion of trust and in the slow poisoning of our shared reality. We can choose to keep patching the wounds, or we can choose to take apart the machine that makes them. The choice should not be hard. We simply need the will to reach for it.