The Pattern Made Explicit

Once is a misfortune but twice is careless, so three times must be deliberate, be it the bombing of British veterans delivering aid to Gaza, or the shooting of Renee Nicole Good by Jonathan Ross of ICE. Donald Trump exalts in that, yet he bemoans the shooting of demonstrators in Iran. In Britain, you can be remanded indefinitely for having silently held up a sign, and permitted to starve to death if, charged with nothing violent and convicted of nothing at all, you dared to protest at such detention. France and Germany are no better.

The photograph of a young woman lighting her cigarette off a burning image of Ali Khamenei is from Canada in 2022. It is of course an expression of contempt for Khamenei, but it is not part of the present events. Yet it has even been copied as a drawing on the front page of The Observer. A video of riots in Los Angeles in 2020 is being circulated as a scene from Tehran on a recent evening. The Islamic Revolutionary Guard Corps ghettoes do famously look like Beverly Hills, but even so. And there is fierce Police brutality, even by the usual standards, in Los Angeles at the moment, against the opponents of the Trumpian Revolutionary Guard Corps, with ICE as its Pasdaran and with all sorts as its Basij, that was subverting the Police, the Armed Forces, and the National Guard. The last Iranian Revolution did not create a democracy, so why should this one? And if Trump wants democracy in Iran, then that is the only country in the Middle East where he does. Herewith, the National Security Strategy.

What Trump wants in and from Iran is, of course, its oil and gas. As he has stated entirely matter-of-factly, the same is true in Venezuela, where there has accordingly been no regime change. The United States has simply staged one of its good, old-fashioned coups within a regime. The new President is a staunch Chavista, but she will play the game, as she had obviously made the arrangements to do. Although it is unclear what game that will be, since the oil companies have decided that Venezuela was too expensive and unstable an investment.

Trump should consider all of that while he eyed Greenland’s rare earths. Since 1951, the United States has been able to have as many military bases as it pleased in Greenland. There is only one left, but it is the Pituffik Space Base, part of the United States Space Force that Trump created and which is therefore especially close to his heart. A Danish flag is flown there, but Pituffik is 930 miles from Nuuk, with no regular flights and with even the irregular ones taking anything up to four and a half hours, often with stops. Might the US be offered Pituffik, although strictly that and nowhere else such as Qaanaaq, as a sovereign base, perhaps with its name changed to Trump?

Our Greenland is Chagos. The Chagos case turns out not to be on any case list at the High Court tomorrow. The House of Lords should either resolve not to consider Third Reading prior to that judgment, or, and preferably, refuse Third Reading as in any event it should. Will Peter Mandelson be there? There would be nothing to stop him.

That it is the former Prince Andrew but not the former Lord Mandelson tells us exactly where power lies. Mandelson was still successfully soliciting donations from Jeffrey Epstein when Epstein was in prison after having been convicted of soliciting prostitution from a minor. The photographs of the two of them with scantily clad young women make Mandelson a liar, and a surprisingly inept one at that. Every other Peer should go on strike as soon as he set foot in the place, bringing the parliamentary process to a standstill. Mandelson is still a member of the Labour Party, so why is anyone else?

Mandelson’s continuing formal and informal role is a far greater threat to our youth than anything on social media. Like banning smartphones at the table as we were banned from having personal stereos, it would be the simplest thing to ban smartphones from schools, which have always called home, or been called, when necessary. But banning social media for under-16s would deny them the formative experience of their generation internationally, together with any ideology other than that of the schools and of the official media. The people who want it in Britain want to lower the voting age to 16, having already raised the school leaving age to 18.

Sexualised images of children are not free speech, nor are non-consensual sexualised images of adults, the United States has no qualms about applying its legislation extraterritorially, and any company doing business here has to abide by our laws when doing so. All of that said, the threat to the X-Twitter is pursuant to the Online Safety Act that was passed when Kemi Badenoch was Secretary of State for Business and Trade, not the immediately responsible Department but not the furthest removed, and when the Conservative Party still delighted in the membership of several people who were now prominent in Reform UK, including Nadine Dorries, the Secretary of State who introduced the Bill, at the committee stage of which the evidence of Hope Not Hate was given by Liron Woodcock-Velleman. Moreover, as Paul Knaggs writes:

When does legitimate regulation become political vendetta? This question sits at the heart of the British government’s extraordinary threat to ban X, a platform used by millions, over failures that plague the entire internet. The answer, it seems, lies not in the technology itself but in who controls it and what they have said about power, especially regarding Censorship.

Grok, the AI chatbot developed by Elon Musk’s xAI and integrated into X, has become a factory for nonconsensual sexual imagery. Research by deepfake analyst Genevieve Oh found the tool churning out approximately 6,700 sexually explicit images per hour during a 24-hour period in January 2026. Users discovered they could circumvent safeguards with absurdly simple prompts, instructing the AI to “undress” uploaded photographs or place subjects in “invisible bikinis.” The Internet Watch Foundation confirmed that criminal imagery of children, apparently created using Grok, had appeared on the platform.

This is appalling. It is unlawful. It demands immediate action. No argument there.

The implications of Censorship in this context are profound, raising questions about freedom of expression and the responsibilities of tech companies.

But here is where the story becomes less comfortable for those who imagine themselves on the side of righteousness. The scale of Grok’s output, while genuinely horrifying, exists within a vastly larger ecosystem of exploitation that operates with minimal government interference. At least 85 dedicated “nudify” websites and applications currently operate online, many accessible through simple Google searches. They advertise on Meta’s platforms. They charge between five and thirty dollars per month. They process millions of images annually. Research published in late 2025 found these sites collectively produced approximately 79 similar images per hour, meaning Grok’s output is roughly 85 times higher than other top services combined.

Yet when Meta’s Instagram and Facebook were found hosting hundreds of advertisements for these nudify tools in mid-2025, promoting the ability to “see anyone naked” and directing users to apps available through Apple’s App Store and Google Play, the government’s response was measured. Regulatory. Proportionate. Meta removed the ads after CBS News investigation. Life moved on. No minister stood before cameras declaring the platforms “disgraceful” or “not to be tolerated.” No threats of total bans. No invocation of national emergency powers.

The pattern holds across platforms. Snapchat, according to multiple studies, remains the preferred distribution channel for non-consensual intimate imagery among teenagers. The reasons are structural: ephemeral messaging, screenshot notifications easily circumvented, and a user base skewing young. DeepSwap, an AI face-swapping service operating from addresses variously listed as Dublin and Hong Kong, was used by a Minneapolis man to create sexualized deepfakes of more than 80 women using photos scraped from Facebook. The discovery sparked local legislative efforts but no existential threat to Meta, DeepSwap’s hosting providers, or the payment processors enabling the transaction.

This is not an argument for inaction. It is an observation about selective enforcement. When the state possesses the power to effectively ban a communications platform used by millions, the principle guiding that power matters more than the immediate target. Either the creation of nonconsensual sexual imagery is an emergency demanding sweeping action across the entire digital ecosystem, or it is a serious problem requiring targeted, proportionate regulation. It cannot be both, depending on who owns the platform in question.

Here we have the arithmetic of enforcement. Technology Secretary Liz Kendall told Parliament that Ofcom should provide updates “in days, not weeks” regarding potential action against X. She confirmed the government’s willingness to back Ofcom in obtaining court orders that would compel internet service providers to block the platform entirely. This represents the nuclear option under the Online Safety Act, deployed precisely six times since the law’s passage. The speed and certainty with which ministers discussed this possibility, the public nature of their threats, and the apocalyptic language employed all suggest something beyond routine regulatory enforcement.

Prime Minister Keir Starmer called the situation “disgraceful,” “disgusting,” and “not to be tolerated,” telling Greatest Hits Radio that he wanted “all options on the table.” When X limited Grok’s image generation to paying subscribers who must provide verified identity information, Downing Street described this as “insulting to victims of misogyny and sexual violence,” claiming it “simply turns an AI feature that allows the creation of unlawful images into a premium service.”

This framing is revealing. Restricting access to verified, paid users who can be identified and prosecuted is not a perfect solution, but it is precisely the kind of friction-based intervention that experts recommend for reducing harm from generative AI tools. It is also considerably more robust than measures implemented by most nudify services, which typically require only an email address and payment information from providers with minimal identity verification. Yet the government dismissed it as insufficient theatre.

The timing compounds the suspicion. Morgan McSweeney, Starmer’s Chief of Staff and the architect of his rise to power, was a founding board member of the Center for Countering Digital Hate, an organisation established in 2018 as Brixton Endeavours Limited before rebranding. McSweeney left the CCDH board in April 2020, two days after Starmer became Labour leader, to take up his current role. The organisation’s founding was funded through McSweeney’s Labour Together think tank, a vehicle explicitly designed, in his own words, to “move the Labour party from the hard left” by undermining Jeremy Corbyn’s leadership. 

Documents leaked to journalists Paul Thacker and Matt Taibbi in October 2024 revealed internal CCDH priorities for early 2024. Among them: “Kill Musk’s Twitter.” The phrasing appeared in multiple internal documents. CCDH CEO Imran Ahmed later characterised this as shorthand for pressuring advertisers to withdraw support and pushing for regulatory action that would damage X’s business model. In December 2025, the US State Department imposed visa sanctions on Ahmed for what Secretary of State Marco Rubio called “organised efforts to coerce American platforms to censor, demonetise, and suppress American speech.”

CCDH’s work has targeted social media platforms for content moderation failures, focusing particularly on vaccine misinformation during the COVID-19 pandemic and climate change denial. It has also campaigned against specific individuals and media organisations, successfully pressing for the deplatforming of conspiracy theorist David Icke and lobbying advertisers to withdraw from outlets including Breitbart, The Daily Wire, and Zero Hedge. According to Paul Holden’s investigation in The Fraud: Keir Starmer, Morgan McSweeney and the Crisis of British Democracy, CCDH and its sister campaign Stop Funding Fake News (now hosted by CCDH) were “incubated using resources” from Labour Together. Holden notes that while both organisations “represent themselves as non-partisan campaigns,” they have disproportionately targeted left-wing media supportive of Corbyn, particularly The Canary, alongside right-wing outlets.

The connections between McSweeney, Labour Together, CCDH, and Starmer’s government are matters of public record, not conspiracy. They demonstrate a pattern of coordinated action across ostensibly independent organisations to shape political outcomes through pressure on digital platforms and media organisations. That McSweeney resigned from CCDH in 2020 does not erase this history. It contextualises the present.

The question is not whether Grok’s failures warrant serious regulatory attention. They do. The question is whether the intensity, speed, and threatened scope of government action reflects genuine concern about online safety or something more calculated. When ministers who showed limited urgency about nudify apps on Meta’s platforms suddenly discover an appetite for emergency measures against X, when they dismiss meaningful safety improvements as “insulting,” when the Prime Minister’s Chief of Staff has documented historical connections to organisations explicitly committed to damaging X’s business model, reasonable people may wonder about proportionality.

This is where defenders of the government’s position will argue that context matters less than outcome, that quibbling over selective enforcement distracts from the urgent need to protect victims. They are wrong. The rule of law depends entirely on consistent application. Regulatory power that bends toward political convenience today will bend again tomorrow, potentially toward targets these same defenders consider illegitimate. Once you have established that threatening to ban major communications platforms is an acceptable response to failures that pervade the internet, you have normalised a tool that will not remain in the hands of those you trust.

Free speech is not an abstract principle. It is the mechanism through which democratic societies identify and correct abuses of power. Platforms matter because they provide the infrastructure for public conversation, however imperfect. When states acquire the practical ability to silence platforms that displease them, and when they exercise that ability selectively based on ownership and editorial stance, the formal existence of free speech protections becomes decoration. The substance has already been compromised.

Regulation is necessary. The Online Safety Act provides a framework for holding platforms accountable for illegal content. Ofcom has legitimate powers to investigate, fine, and compel changes. The creation of nonconsensual sexual imagery, whether through traditional photography or AI manipulation, should be criminalised comprehensively, with particular severity for images depicting children. Platforms should face meaningful consequences for systemic failures to prevent and remove such content. Payment processors and advertising networks that enable the economics of exploitation should be subject to coordinated enforcement action.

But banning an entire platform, used for political discourse, journalism, emergency communication, and countless other legitimate purposes, because its AI chatbot shares a failure common to dozens of other services, while those other services face no comparable threat, is not regulation. It is power dressed in the language of protection.

The UK is heading down a very dark path. Over 12,000 people were arrested in 2023 alone under communications offences for social media posts. Thirty arrests per day for messages causing “annoyance,” “inconvenience,” or “anxiety.”

The Labour government did not enter office promising to restore integrity to British politics. That was the marketing. The reality, evident within months, tells a different story: systematic suppression of dissent through mechanisms that would make authoritarian regimes envious.

Multiple journalists detained under terrorism legislation for reporting on Palestine. Justice Secretary David Lammy proposing to strip the right to jury trial from 75% of Crown Court defendants, limiting this ancient safeguard to only murder, rape, and manslaughter. A former MP and leader of a British political party was arrested and detained at Gatwick Airport. George Galloway and his wife were detained by counter-terrorism police, a stark reminder that in Starmer’s Britain, the machinery of the surveillance state operates with chilling efficiency against political opponents as well as journalists and even comedians.

The former MP was stopped by officers under Schedule 3 of the Counter-Terrorism and Border Security Act after arriving from Moscow via Abu Dhabi. The legislation allows authorities to stop, question, search and detain individuals at borders to determine if they have engaged in “hostile activity,” a conveniently elastic definition that can apparently stretch to cover any political position the government finds inconvenient.

The threat to ban X is not an aberration. It is the pattern made explicit. 

Human Rights Watch released its damning assessment in January 2026: the Labour government has taken “a deeply alarming direction on protest rights” and is adopting “protest-control tactics imposed in countries where democratic safeguards are collapsing.” The government not only refused to repeal Tory anti-protest legislation described as “deeply troubling” by the UN High Commissioner for Human Rights, it defended it in court and expanded it. Five Just Stop Oil activists received prison sentences of two to five years merely for joining a Zoom call to plan a protest. When courts called these sentences “manifestly excessive,” they reduced them marginally. The message was clear: planning peaceful protest is now treated as conspiracy. 

Journalist Richard Medhurst was arrested at Heathrow under Section 12 of the Terrorism Act in August 2024 for “expressing an opinion or belief that is supportive of a proscribed organisation.” He was held for nearly 24 hours, his journalistic equipment seized, access to a lawyer denied. The charge related to his coverage of Palestine. In October 2024, counter-terrorism police raided the home of Asa Winstanley, associate editor of The Electronic Intifada, seizing his devices under Operation Incessantness. Sarah Wilkinson was arrested in her home. The International Federation of Journalists condemned these arrests as “apparent misuse of anti-terror legislation.” Not a single mainstream British newspaper covered the raid on Winstanley’s home. The silence was instructive…

The government proscribed Palestine Action in July 2025. Since then, at least 2,300 people have been arrested for showing support. The organisation’s crime? Direct action against arms manufacturers supplying weapons used in Gaza. Sixteen members remain in custody, most on remand without conviction. This is the new Britain: expressing solidarity with proscribed organisations, planning protests, posting messages that cause “anxiety,” standing silently with signs outside abortion clinics, reporting on occupied territories. All now punishable by arrest, prosecution, imprisonment.

Freedom House downgraded the UK’s internet freedom score from 5 to 4 in 2025, citing the “proliferation of criminal charges, arrests, and convictions concerning online speech, including speech protected under international human rights standards.”

This is not hyperbole. It is documentation. Morgan McSweeney, the architect of Starmer’s rise and now his Chief of Staff, co-founded the Center for Countering Digital Hate, an organisation whose leaked internal documents listed “Kill Musk’s Twitter” as a priority. McSweeney’s Labour Together think tank, designed explicitly to destroy the Corbyn left, pioneered the tactics now deployed across the British state: suppression disguised as safety, censorship marketed as responsibility, authoritarian control rebranded as protecting the vulnerable. The same methods that targeted left-wing media outlets like The Canary now target X. The same apparatus that undermined democratic socialism within the Labour Party now undermines democratic discourse across Britain.

The threat to ban X is personal, political, and precedent. It represents the fusion of state power with factional interest, the subordination of civil liberties to political convenience, the transformation of regulatory authority into an instrument of control. If it succeeds, if a government can effectively silence a major communications platform because its owner criticises power and its AI tool shares failures common to the entire industry, then the principle is established. The tool exists. It will be used again. Against different targets, under different pretexts, toward the same end: limiting the spectrum of acceptable opinion while permitting lively debate within those ever-narrowing bounds.

We are not watching the first serious test of whether Labour’s government believes its own rhetoric about democratic norms. We are watching the answer. The arrests speak. The imprisoned protesters speak. The detained journalists speak. The restricted juries speak. The threatened platforms speak. They say, with clarity that should terrify anyone who values liberty: this Labour government is Labour in name only. It has become the instrument of precisely the authoritarian project it once claimed to oppose. And it shows no sign of stopping until every avenue of meaningful dissent has been legislated, prosecuted, or intimidated into silence.

When governments treat peaceful protest as conspiracy, journalism as terrorism, and platforms as enemies, democracy has already fallen. Only the paperwork remains.