‘Lean Into Who You Are’: Viner on the Guardian’s Bet Against AI
At WAN-IFRA’s World News Media Congress, the Guardian’s editor laid out the defences her newsroom has spent a decade building. Then she named the one threat they may not stop.
The Guardian may be the best-protected newsroom in the English-speaking world. No proprietor to please, no shareholders demanding a dividend, a trust whose sole purpose is to keep the paper alive in perpetuity, and a reader-revenue base most publishers can only envy. Katharine Viner spent an hour explaining how those defences were built. Only once did the harder question surface, and she answered it candidly: whether any of it protects against the threat now coming.
The session sat within a framework set by Stig Kirk Ørskov, WAN-IFRA’s chief executive, who opened the congress. Press freedom, he argued, is a human good as much as a democratic one: line up the World Press Freedom Index against measures of public trust and wellbeing and the correlations hold. His precondition was blunt. Journalism cannot stay free if it is not financially sustainable, which turns a fair exchange of value between publishers and the AI firms ingesting their work into an existential question rather than a commercial nicety. That precondition framed everything Viner went on to say.
Viner has edited the paper for eleven years to the day, as her interviewer, AFP global news director Phil Chetwynd, reminded her. She took the chair a year before Brexit and a year before the first Trump term. “It’s been a decade of absolute turmoil,” she said, then widened the frame past media economics: more people now live under autocracies than democracies, the climate crisis is accelerating, and the West is logging its own figures on loneliness, poor mental health and young people locked out of housing and entry-level work. Her instinct throughout was to place journalists inside that picture rather than above it. “It’s important for journalists to see ourselves as citizens, not as a sort of different group of people sitting on top of society.”
The third crisis
Viner borrowed her diagnosis from the novelist Naomi Alderman: the information flood is the third great rupture in how humans handle knowledge, after the invention of writing and the advent of the printing press. The symptom is saturation, not scarcity. “We are flooded with information, flooded with facts,” she said, and the disorientation that follows, things you assumed true turning out false, things you dismissed as fake turning out real, is compounded by the sheer volume of deliberate manipulation. She cited the claim that most of what is now online is synthetic, and therefore probably fake. Chetwynd supplied a live example from the Iran crisis: the White House’s X account and the Iranian government’s, pumping out fabricated material in parallel, a point Viner took up at once. “How can you blame people for being confused about what they see?”
The conclusion is as commercial as it is editorial. Where the ground shifts underfoot, a recognised source of trust becomes a scarce asset. Facts alone do not close the gap. “You can’t just slam them on the table and say, well, here are the facts, because people will always provide another set of facts.” The Guardian’s answer is to wrap reporting in story and context, and to meet readers in the format they use rather than lecturing them: a reader who only watches video gets nothing from a 4,000-word essay. News avoidance, the industry’s running anxiety, is not something Viner says she sees in the Guardian’s own numbers.
Refusing the bait
The discipline shows most clearly in how the paper covers Trump. Viner traced the policy to a single morning in 2016, when she woke to find the Guardian, the BBC and The Times had each separately fact-checked the same speech overnight. “The whole world is going to spend the next four years fact-checking this guy. That is a waste of our collective time.” The Guardian’s response was to report less of what Trump says and more of what it does to people’s lives: “people as well as power,” a phrase she returned to repeatedly. The alternative is familiar to anyone who covered the first term: a newsroom reduced to a distribution channel, fed fresh lines morning, noon and night. Chetwynd recalled AFP’s Washington desk dismissing the bleach-injection remark as absurd, then having to cover it once it became consequential. Ignoring him is not an option either, which leaves the newsroom to weigh each remark on whether it changes lives or merely makes noise.
Getting off the platforms
The same refusal to be used drives the Guardian’s social-media policy. The paper no longer expects its journalists to be active on social platforms, a reversal of the era when it pushed them on. Viner’s reasoning is partly about judgment: once a reporter is composing for an audience of critics, “you start hearing voices in your head telling you should have done it that way, and then you make bad decisions.” The Guardian has left X altogether, on the grounds that staying “just feeds a far-right algorithm to make somebody else rich and change politics.” No individual is compelled to follow, but the direction is set. She has also raised spending on physical protection for reporters, citing the number of media workers killed in Gaza as evidence that the press vest no longer offers the cover the industry once assumed, “a sign of the times.”
The model that holds
Behind all of it sits the Scott Trust. No proprietor extracting cash or influence, no shareholders pressing for margin. Viner is blunt that this is the best arrangement an editor could ask for and offers two illustrations of what it buys. The first was a libel case the paper fought for six weeks in the High Court and won comprehensively, with the trust’s full backing, a result she believes has measurably improved the climate for investigative reporting in Britain, where the law tilts towards the claimant. The second was the reader-revenue model itself. When the Guardian began asking readers in 2016 to pay for journalism they could previously get for free, the idea was ridiculed across the British press and within the building. The trust backed it, and readers grasped it faster than the critics had.
The figures Viner cited make the case. Recurring reader revenue has gone from nothing a decade ago to roughly £125m last year across print and digital, she said, before one-off contributions; the digital share alone rose 22% to £107m in the year to March 2025. The model has also gone global: by her account, 83% of the Guardian’s revenue now comes from outside the UK, almost none of which existed a decade ago. What makes the model durable, in her account, is that it is not a transaction. “You’re not a consumer in the traditional sense; you are more part of our community. It’s not a transaction, it’s a choice.” She claims this makes it more resilient than a paywall, while conceding the obvious limit: the share of regular readers who pay stays small. The current work is adding transactional on ramps for people who prefer a straight exchange: the Feast recipe app, the Guardian Weekly, the news app, while keeping the website free to all. That last point she frames as a matter of mission. With paid news becoming the default, a free open site is a genuine differentiator.
The one fear she names
Viner is not against the technology itself and offered a concrete case where it earned its place. Working with University College London, the Guardian analysed a century of rhetoric on immigration in the British Parliament and found that the language of the past five years had been more aggressive and more anti-immigration than at almost any point in the preceding hundred. The study, she said, would not have been possible without AI. The journalists most excited by the tools, she added, tend to be the ones who push the rest of the newsroom forward.
For all that, Viner named a single threat without hedging. AI’s capacity to sever the link between the Guardian and its readers, the disintermediation AG Sulzberger had warned about in the congress keynote, is what worries her most. Everything the paper has built rests on a direct relationship with its audience. An answer engine that satisfies the reader before they reach the site dissolves that relationship at the root. Her response is to lean harder into the channels she controls and the community she has spent a decade cultivating.
Here the two newsrooms part company. Where Sulzberger opened his keynote with a call to arms, Viner’s posture is quieter and more collective. On litigation and licensing, she was deliberately vague. The Guardian is party to some class actions, she confirmed, though not at his scale, and it has struck deals to secure compensation for its work. “You do different things at different moments with different organisations.”
The collective answer is SPUR, the Standards for Publisher Usage Rights coalition. Launched in February by the Guardian alongside the BBC, the Financial Times, Sky News and the Telegraph, it is a non-profit set up to build shared, machine-readable standards and licensing frameworks for how AI systems use journalism. The aim is to let developers reach reliable reporting through paid-for channels while publishers keep practical control of their content and are paid fairly for it. Its founders have written to publishers and broadcasters worldwide inviting them to join, and Viner expected an imminent announcement of further global members. This is the publisher-built version of the value exchange Ørskov had described, where Sulzberger turned to litigation, Viner has turned to the coalition. It arrives as the UK government weighs its own position on AI and copyright. Viner was sharper on that: “Governments around the world seem so desperate for growth that they seem to think the words AI equals growth.” The creative industries, she suggested, keep getting left out of that sum.
She was strikingly un-snobbish about where audiences now find news. Influencers, she argued, have done what newsrooms have struggled with: built close relationships with their audiences and learned their platforms cold. The opening she sees is to fuse that with editorial standards: a Guardian news influencer whose information readers can trust, delivered in a form that fits the platform. The format mix is already moving: the Guardian publishes two novels’ worth of words every 24 hours, a proportion she admitted is “out of whack,” strong in audio, decent in video, with room to do far more for younger audiences.
Asked for one piece of advice for editors, Viner gave the line that doubles as the Guardian’s entire strategy: lean into who you are. “The values of who we are is what differentiates us in this moment.” For a small publisher, she added, a community built around a clear identity is a sound basis for a business.
It is good advice. The Guardian’s defences, the trust, the reader community, the editorial discipline, the open site, were built against the threats of the past fifteen years: proprietorial meddling, the collapse of display advertising, the erosion of public trust. Against those they are formidable. Whether a community survives when AI bricks up the front doors the audience walked through, search and social and the open web, is the bet Viner is now placing the institution on. She knows it. With Ørskov’s precondition still unmet, the honest reading of the hour is that the best-defended newsroom in the language does not yet know the answer, and neither does anyone else.