Standing before the United Nations General Assembly in New York, the UK prime minister delivered what felt like a warning and a promise. With the world’s attention fixed on war, climate and inequality, he insisted that the next great battle would be fought not only on physical battlefields but in the information we consume and the algorithms that shape our thinking. The United Kingdom, he argued, wants to be at the forefront of setting global safety standards for artificial intelligence—an ethical framework that can keep pace with a technology moving faster than laws, governments and public understanding. He described AI as a force of enormous promise, capable of transforming medicine, education and industry, but also as a force that, without careful guardrails, could deepen existing fractures in democracies. The speech was his first major international address since taking office, and it was notable for its plain-spoken urgency. He spoke not as a technologist but as a political leader worried about what technology could do to people’s lives, and he refused to sugar-coat the scale of the challenge. Rather than offering vague optimism about innovation, he used the platform to name the dangers and call for collective action. The UK, he said, was ready to act as an honest broker—a nation that could convene allies, mediate between competing interests and lay down a foundation of rules and norms before it is too late. That phrase, “honest broker,” was chosen carefully; in an age of suspicion and fragmentation, he was asserting that Britain could be trusted to balance technological progress with human rights and democratic resilience. Yet the audience was also reminded that trust alone is not enough. The prime minister warned that existing threats—from cyber-attacks to disinformation—will not simply continue; they will multiply as artificial intelligence becomes more sophisticated. The speech was an attempt to wake the world from its complacency, to transform AI from a distant abstract debate into an urgent matter of national security and social cohesion. It was also, in many ways, a declaration of national purpose: the UK does not want merely to compete in the AI race; it wants to help write the rules by which that race is run.
At the heart of the announcement sat a new institution: the National Centre for Information Defence. Its mission is to detect, expose and counter AI-assisted disinformation campaigns operated by hostile states. The prime minister described the centre as a shield for the truth, a way to “stem the poisonous tide” of lies aimed at destabilising democratic societies. The language was deliberately strong. This is not a minor bureaucratic addition; it is a recognition that information has become a weapon of war. This is a battle for hearts and minds in the most literal sense, and the new centre is intended to meet it with speed and agility. For years, governments have known that foreign actors use social media to sow confusion, inflame divisions and undermine confidence in everything from elections to public health advice. What the UK is now acknowledging is that AI makes these attacks cheaper, faster and far more difficult to trace. A single operator can generate thousands of realistic fake articles, videos and social media posts in moments. The new centre will work with tech companies, intelligence agencies and international partners to track these campaigns in real time and alert the public before lies can take root. But the centre’s purpose is not purely defensive. By making disinformation visible, it also aims to restore something more fragile: people’s trust in what they read, watch and share. The human dimension is easy to overlook when discussing algorithms and bots, but the prime minister brought it back into focus. Disinformation does not simply confuse people; it makes them feel isolated, angry and fearful in their own homes. It turns neighbours against one another, makes communities suspicious of institutions, and creates a murky atmosphere in which facts become opinions and lies become equally respectable choices. The new centre, he suggested, is a response to that atmosphere—an attempt to give citizens the tools to recognise manipulation and to hold powerful actors accountable. The name “Information Defence” matters: it implies that truthful communication must be guarded, that freedom from manipulation is a basic right, and that every democratic country needs a defence system not just for its borders but for its public mind.
The prime minister was explicit about the source of the most serious current threat. The details were specific enough to feel credible; this was not a vague accusation but a documented pattern of hostile activity. Russia, he said, is running a vast and insidious disinformation campaign against the United Kingdom and other Western countries. The Kremlin spends around £1.3bn each year on manipulating information—money used to create an entire ecosystem of deceit, from bots and fake websites to fabricated newspaper articles. Perhaps most disturbingly, Russian operatives have forged the branding of 28 British organisations, including universities and the BBC, lending their false content the appearance of trusted sources. This is not random internet trolling; it is a structured, well-resourced effort to distort how people perceive their own society. The prime minister described it as an insidious campaign that reaches into people’s homes and twists how they feel about their country and community. Its aim, he said, is to create a narrative of decline, to stoke division and to sow despair. The result is a slow poisoning of public life: people begin to believe that their institutions are corrupt, that their neighbours are enemies, and that there is no common ground left. The mention of the BBC is especially striking, because the BBC is one of the most trusted news organisations in the world. If people cannot rely on the broadcaster’s name, if even that symbol of impartiality can be stolen and weaponised, then the entire idea of a shared reality is at risk. The prime minister also warned about cyber-attacks against British companies and institutions, hinting that these digital assaults are not separate from disinformation but part of the same strategy. First, undermine confidence in the truth; then, attack the systems that keep society running. Together, these actions create a mood of vulnerability and cynicism. And now, with artificial intelligence, the scale of the threat is about to expand dramatically. A country that used to need a large team of human operatives to produce and spread fake stories can now rely on AI to do the work around the clock, in dozens of languages, with almost no cost. The challenge is no longer simply to debunk individual lies, but to defend the very concept of reality from being overwhelmed.
To illustrate the stakes, the prime minister pointed to warnings from those closest to the technology. Evan Hubinger, a safety researcher at the AI company Anthropic, recently cautioned that the technology is advancing so quickly that there is a chance it could “kill all humans” within the next decade. The phrase sounds extreme, but it comes from someone who spends his working life studying the behaviour of advanced AI systems. The point is not that AI is an evil monster waiting to escape; it is that powerful systems are being built faster than our ability to understand their capabilities and control their risks. If a machine can be given goals and the ability to pursue them, humans may not always be able to predict what it will do. Meanwhile, closer to everyday concerns, the UK’s AI minister, Kanishka Narayan, acknowledged that “people are right to worry” about the impact of AI on jobs. Whole professions could be transformed or displaced in a matter of years, and while new jobs will appear, they may not favour the same people or the same communities. These two warnings describe the spectrum of anxiety: existential catastrophe at one end, and the quieter, more familiar fear of losing one’s livelihood at the other. Between them lies a vast territory of other concerns—bias in automated decisions, invasions of privacy, the use of facial recognition in policing, the manipulation of elections, the destruction of creative professions, the concentration of power in a handful of corporations. The prime minister did not pretend to have all the answers. Instead, he argued that these are not problems for individual nations to solve alone. AI is inherently global. The data crosses borders, the models are trained in one country and used in another, and the consequences of failure will be felt everywhere. That is why the UK wants to lead the creation of international safety standards—rules that can govern how AI systems are designed, tested and deployed. Without such standards, states will react in chaotic and contradictory ways: some will create too little regulation, allowing dangerous systems to spread; others will create too much, choking off innovation. The government is therefore trying to prepare the public for a future of constant adaptation, while also pressing for safeguards that reduce the likelihood of catastrophe. The hope is that a shared framework can steer between panic and paralysis.
The UK’s claim to leadership rests on a particular role: that of an honest broker. It is a claim that will need to be tested, but the moral authority of honesty is not something that can be bought with a budget. The prime minister used the phrase to signal that Britain is not trying to sell a particular technology or advantage over rivals; it is trying to build consensus. In practical terms, this means convening governments, scientists, industry leaders and civil society groups to agree on what constitutes safe and responsible AI. It means pushing for transparency so that citizens can understand how decisions are made by machines. It means creating mechanisms for accountability when systems fail or cause harm. And it means ensuring that the benefits of AI are not hoarded by a few wealthy countries and corporations, but shared fairly across the world. This is an ambitious agenda, and there are reasons to be sceptical. Britain is not the largest developer of AI technology; that title belongs to the United States and China. Nor is it the largest regulator; the European Union has already advanced its own ambitious AI Act. So why should the world listen to the UK? Perhaps because the country occupies a useful middle ground. It is deeply connected to American tech and intelligence communities, but it retains independent institutions and a strong public broadcaster. It is a democracy with parliamentary traditions and a rule of law, and it has historically played a role as a bridge between Europe and the wider Anglosphere. The claim to honest-broker status is also a rebuttal to the politics of resentment and disinformation. In a world where rival powers are using AI to lie to each other’s populations, Britain is saying that there is still a place for a country that tells the truth, even when the truth is uncomfortable. Leadership is not only about having the most advanced technology; it is about having the credibility and imagination to build an order that people can trust. The prime minister’s speech was therefore not a technical presentation but a political argument: that safety standards for AI are also standards for civilised behaviour among nations. Without them, the race to dominate AI will deepen mistrust, accelerate arms races and leave ordinary people caught in the crossfire of information warfare.
In the end, the message to the United Nations General Assembly was simple: the future is being written now, and it will be written either by accident or by design. The UK’s announcement of a new centre to defend truth is an acknowledgement that the battle for public trust has already begun. The warnings about Russia’s disinformation campaign, the concern about AI’s existential dangers and the anxiety about jobs all point in the same direction—the need for a more mature, more responsible approach to technology. None of this is inevitable. AI is a tool, not a fate. It can be shaped by human choices and human values. The prime minister’s speech sought to remind the world that those values matter more than ever: honesty, solidarity, humility and the courage to confront uncomfortable truths. Establishing global safety standards will be difficult. The interests of major powers are deeply aligned with their own technological dominance. But the very fact that the debate is happening at the United Nations is a sign of progress. Ten years ago, world leaders were not talking about AI safety at the General Assembly. Now they are. The problem has been named, and that is the first step toward solving it. For ordinary people, the speech was a reminder that these issues are not abstract. They will shape the news we see, the jobs we hold, the way we vote and the way we trust one another. A world in which AI is used to spread lies and division is a world in which everyone becomes poorer, more frightened, more isolated. A world in which AI is governed by shared standards is a world in which technology can still be a source of hope. The UK has placed its flag on that hopeful ground. Whether other nations follow will depend on their willingness to see the truth as something worth defending. Perhaps that is the real meaning of leadership: not to promise perfect answers, but to create the conditions in which better answers can be found together. The prime minister ended his address not with a threat but with an invitation: to build a future where power is tempered by responsibility, and where innovation serves the common good. That is a message worth hearing, and perhaps, at last, it is a message whose time has come.