Welcome back to the Citizens Understand.

For years, campaigners, researchers and organisations have been pointing to the same pattern. New technologies are rolled out at speed, harm follows, and accountability arrives too late, if it comes at all. Social media came before it. Now AI is doing the same, from the rise of deepfake material to the proliferation of ‘nudification’ apps aiding child abuse.

A new report brings this into sharp focus. The first comprehensive study into how AI chatbots are implicated in violence against women and girls (VAWG) lays out, in grim detail, how this harm is being built into the systems themselves, as untested, rapidly deployed technologies begin to re-embed misogyny into everyday digital life.

In this edition of the Citizens Understand, we dig into the report to show how this is already playing out and what can be done about it.

📝 What the report says

Women are already more wary of AI than men, using services like ChatGPT far less frequently. This caution is justified: the normalisation of abuse is now embedded in chatbot design. In their landmark report, Invisible No More, Clare McGlynn and a team of academics identify a new typology of violence against women and girls (VAWG) that proves these fears well-founded.

As McGlynn and her colleagues establish, we are facing four distinct new forms of violence. There is chatbot-driven abuse, where the system itself generates harmful content. Chatbot-enabled abuse, where it helps users carry it out. Chatbot-simulated abuse, where it takes part in abusive roleplay. And chatbot-normalising abuse, where it frames that behaviour as acceptable, trivial or even desirable.

The scale of this ‘chatbot-driven’ harm is staggering. Take Grok, which quickly became an industrial-scale machine for the production of sexual abuse material. In just 11 days, the system generated an estimated 3 million sexualised images of women and girls.

The consequences are already shaping how young people understand harm. Sophie Lennox of Everyone’s Invited told us: “AI chatbots and deepfake technology is becoming increasingly normalised within the classroom.

“At Everyone’s Invited we are seeing this in realtime. In one of our sessions, part of our education programme reaching 90,000 young people, a boy had been sexually deepfaked by his peers.

“Almost all of the students in the class thought that what happened to him was not harmful, and crucially, they thought it was normal.”

The unprecedented mainstreaming of nudification technology forced public attention onto the issue. While this undoubtedly shifted awareness, it also pointed to something much bigger. A growing ecosystem of tools designed to produce, refine and distribute new forms of abuse.

The report’s most significant finding lies in what it calls chatbot-simulated violence. This is where harm emerges through interactive roleplay, with chatbots actively co-producing abusive, gendered scripts in real time.

Chub AI is one example of chatbot-simulated violence cited in the report, a platform that recorded 11.3 million visits in January this year. Users can select characters and scenarios to interact with. In her investigation, Laura Bates found that Chub AI offered users access to scenarios including a so-called “brothel”, staffed by girls under 15, for sexual roleplay. This comes alongside reports that the ‘most popular scenarios on Chub involve text-based child sexual abuse’. It is now banned in the UK and Australia.

A 2025 report by Graphika looked at another platform, Character.AI, alongside platforms including Spicy Chat, Chub AI, CrushOn.AI and Janitor AI. It identified more than 10,000 chatbots explicitly labelled as sexualised “minor-presenting” personas or engaging in roleplay featuring sexualised minors. Character.AI has since introduced restrictions for under-18s following legal action brought by the parents of Sewell Setzer, who took his own life after months of engagement with the platform.

But those safeguards do not hold up in practice. Everyone’s Invited has been investigating Character.AI in recent months and has found its age verification was easily bypassed by entering any birth date, with no further checks in place.

Equally, harmful sexualised content appeared almost immediately in some interactions, with numerous characters referencing rape and controlling behaviour.

The report also exposes how Character.AI’s very architecture facilitates abuse, built into the system in multiple ways: there are no meaningful rules to block abusive roleplay scenarios; it is designed to be “sycophantic,” prioritising affirmation and continuation over any form of refusal; it has no filtering for sexual violence, coercive dynamics or VAWG-relevant patterns; and, crucially, the platform assigns sole responsibility to users for the co-production of these abusive narratives.

🤷‍♀️ What can be done?

The report provides a vocabulary for a violence that, until now, has been difficult to name. It exposes exactly how these systems are implicated in the production of harm. But naming the problem is only the first step, and it is being outpaced by a global surge in investment. Less than two weeks ago, OpenAI closed a funding round valuing the company at $852 billion, cementing its position as one of the most highly valued private companies in the world.

This has occurred, in part because of a well-oiled Big Tech machine that has spent decades gaslighting the public into believing regulation is impossible. The myth that intervention stifles innovation or breaks progress has been repeated so often it is now accepted as fact. Which is why it is unlikely that change will come from within these companies themselves.

Looking to the government, there are signs of movement. Keir Starmer recently described violence against women and girls as a “national emergency”.

Today (Tuesday 14 April), the Commons will consider amendments to the Crime and Policing Bill, which will ban any AI tool capable of generating deepfake nudes. The Bill introduces a long-overdue “48-hour rule,” forcing platforms to remove non-consensual intimate images within two days of a report.

While these measures tackle the symptoms of this violence, government continues to stall on the source. An amendment by Baroness Beeban Kidron to create a new criminal offence for “unsafe AI chatbots” was rejected. Instead, the government has retreated into the familiar territory of delay, committing only to a report by December on how the Online Safety Act might eventually regulate these systems.

We’re a reader-supported newsletter explaining how power operates and organising to challenge it. Consider becoming a free or paid subscriber.

Subscribe now

McGlynn and her team call the current approach “wholly inadequate”. Reactive responses, shaped after harm has already occurred, are not good enough. What’s needed is a shift in approach that focuses on how this harm is produced in the first place.

The report calls for chatbot-simulated violence to be recognised as an “emerging and distinct category of risk requiring urgent regulatory attention”. That means clear accountability across the AI supply chain, enforceable safety-by-design standards, and proactive oversight to prevent the large-scale reproduction of gender-based violence through chatbot simulations.

McGlynn also proposes a new criminal offence of “dangerous deployment of an AI chatbot,” targeting companies that release products without taking reasonable steps to prevent harm. Crucially, this is not an argument against the technology itself, but for it to be treated like every other sector, where safety is built in from the start.

On what needs to change, Lennox of Everyone’s Invited said: “For too long, survivors have been expected to share their experiences of abuse in order for action to be taken. The recent legal cases against Meta mark an important step towards accountability.

“But this cannot be where it ends. Safety by design must be built into all emerging technologies. Harm must be prevented.”

🧐 Our analysis

Part of the reluctance to act against AI chatbots comes from the idea that these harms are somehow intangible. That what exists online does not carry the same weight. If an image is generated rather than real, how much damage can it really do? Is it enough to justify regulating the entire sector?

When framed that way, I think of the ten and 11-year-olds dropping out of school after being targeted by deepfake sexual abuse. I think of children being lured into ‘romantic and sexual roleplay’ by AI chatbots instead of getting help with their homework, or the quarter of teens turning to these systems for mental health advice. I think of the serial stalker, accused of harassing over a dozen women, who was actively encouraged by ChatGPT to continue terrorizing his victims.

The idea that this technology lacks real-life consequences for women and girls is absurd; but nor will it help men and boys. Companies marketing themselves as “companion who cares” or offering “unbounded connection” merely mask the fact that they profit from a business model rooted in the commodification and exploitation of women. These apps don’t provide a safe outlet; they train all of us to view women as submissive products, an entitlement that is far more likely to escalate real-world harm than prevent it

This report establishes that the harms of this technology aren’t a “future dystopia” or a theoretical concern about a robot uprising. It is devastating lives, right now, part of a wider technological shift that is intensifying and inventing new forms of gender-based violence. How much more evidence is needed before this is taken seriously? How many survivors will it take for tangible solutions to emerge?

Subscribe now

✊ How to fight back

Everyone’s Invited was built on the voices of survivors of rape culture and sexual violence.

👉 Share your story or get involved to support their work to end this culture.
👉 Attend Everyone’s Invited’s event at Marmalade Festival on 24th April, where they will be speaking alongside the Future of Life Institute and activist Megan Garcia on AI harms. Register here.

Write to your MP and ask where they stand. Do they support stronger regulation of AI chatbots, including new offences for unsafe deployment? What are they doing to push back against the ways Big Tech enables violence against women and girls?

🔎 Find your MP.

Pause AI is calling for a pause on advanced AI development until international safety agreements are in place.

👉 Join the growing movement pushing for democratic oversight.

See you next time,
Team Citizens

Share

About The Citizens Understand:
In an era where technology is reshaping democracy faster than laws can keep up and power is increasingly exercised through platforms, The Citizens Understand exists to cut through misinformation and make complex systems legible. If there’s something you’d like to understand, email lillian@the-citizens.com.

For years, governments and defence contractors have predicted that artificial intelligence would transform modern warfare. That future is no longer theoretical.

A new era of conflict is here, one where the speed and scale of military decision-making is shaped by machines, and AI-powered bombing is even more deadly.

If Gaza is ground zero for the Israeli AI war machine, the aggression of the United States and Israel against the people of Iran is a major demonstration of how AI systems function as integrated components in modern military campaigns.

The war has also exposed a growing struggle over who controls this technology. In Washington, a bitter row erupted between the Pentagon and AI company Anthropic after the firm refused to allow its models to be used for some military purposes. Rival tech companies rushed to fill the gap.

In this edition of the Citizens Understand, we examine how artificial intelligence entered the battlefield and what this means for the future.

How AI entered warfare

The shift from consumer AI to military infrastructure didn’t happen overnight. In 2024, the San-Francisco AI company Anthropic began deploying its model, Claude, inside parts of the US Department of Defense and other national security agencies. The aim was simple: speed up the way wars are planned.

Anthropic partnered with spytech firm Palantir - the company embedding its software deep inside Western governments - to make Claude available to US defence and intelligence networks.

Palantir aggregates vast streams of military intelligence and feeds them into AI models like Claude, enabling “faster, more efficient and ultimately more lethal decisions where that’s appropriate,” according to its UK chief Louis Mosley.

Palantir is rapidly embedding itself inside military operations and governments. In the UK alone, its contracts with the state now total at least £670m.

✍️ Sign our petition calling for a full Parliamentary debate on Palantir.

Sign our petition

Anthropic’s tools were approved for use inside classified systems under an agreement first signed with the Pentagon during the Biden administration. But the company insisted on two ‘red lines’.

First, Claude could not be used for mass surveillance of American citizens. Second, it could not control fully autonomous weapons capable of selecting and attacking targets without a human operator. Those limits were later reviewed and accepted by the Trump administration.

At the time, the arrangement was presented as a model for responsible cooperation between Silicon Valley and the government. But it would not last.

Tensions came to a head in January, when reports emerged that the US military had used Claude during an operation to capture Venezuelan president Nicolás Maduro.

When the company began pressing for answers about how its technology had been used in the raid, a spat broke out and soon spilled into public view.

After Anthropic refused to relax its red lines, Donald Trump ordered all federal agencies to stop using the company’s technology, denouncing Anthropic on Truth Social a “radical left woke company”. Defense secretary Pete Hegseth went further, designating the firm a national-security supply-chain risk. The systems remain in use while they are phased out.

Anthropic’s chief executive, Dario Amodei, responded that the company would rather walk away from Pentagon contracts than allow its technology to be deployed in ways that could “undermine, rather than defend, democratic values.”

Shortly after, Sam Altman’s OpenAI–whose president is Trump’s biggest donors–swooped in to take over the $200 million contract. Anthropic’s statement makes clear exactly what OpenAI has now agreed to.

Altman said the deal would include the same restrictions that triggered the standoff with Anthropic: no autonomous weapons and no domestic mass surveillance. That contradiction has not been addressed.

AI helps bomb Iran

Just hours after the Donald Trump administration severed ties with Anthropic, the United States and Israel launched a wave of strikes across Iran - reportedly using that very same company’s AI as part of the operation.

The The Wall Street Journal revealed that US strikes in the Middle East relied on Anthropic’s model Claude, while The Washington Post reported the military had “leveraged the most advanced artificial intelligence it’s ever used in warfare”.

The tools were used to sift through vast volumes of intelligence and help planners decide which targets to strike, a process designed to “shorten the kill chain”.

The opening wave of attacks killed Iran’s supreme leader Ali Khamenei and triggered strikes across multiple provinces. At least 165 people—many of them children—were reported dead after an explosion destroyed a girls’ school in southern Iran. AI’s role in these events is not yet known.

Gaza was the pilot

For more than two years, Israel has used Gaza as the testing ground for the technology now used in Iran.

The most extensively documented of these is a system known as Lavender. An investigation by +972 Magazine revealed Lavender was used to analyse vast intelligence datasets and automatically flag potential targets. At one point, the system reportedly identified around 37,000 Palestinians as suspected militants, operating with an accepted error rate of roughly 10 percent.

Where are the rules?

The International Committee of the Red Cross has repeatedly warned that autonomous and semi-autonomous weapons risk breaching international law. Warfare depends on human judgement, it argues, weighing proportionality, distinguishing civilians from combatants and taking precaution to minimise harm. They are moral and contextual decisions.

Despite these warnings, governments and technology companies continue pushing the technology forward. Part of the reason lies in the venture-capital logic now shaping defence technology, where the same Silicon Valley priorities - speed, scale and competition - now shape how military technology is built and deployed.

Efforts to agree binding rules on autonomous weapons under the Convention on Certain Conventional Weapons have repeatedly stalled. Major military powers, including the United States, have resisted a treaty, opting instead for voluntary guidelines.

Analysis

What we are witnessing may be the first full-scale artificial-intelligence war in history. The clearest expression yet of how the most powerful technologies of the twenty-first century are being folded into the machinery of violence.

The same logic that built recommendation engines and social media algorithms is now being applied to the industrial generation of kill lists. Chatbot systems that entered public life writing emails, summarising documents and telling you what to have for dinner now help determine who lives and who dies.

In simulated war games, AI models given autonomous control over weapons systems opted to use nuclear weapons in 95 per cent of cases. That should shake us to our core.

By design, autonomous and semi-autonomous weapons erode accountability, distributing decisions across complex systems of algorithms, data and operators, making responsibility harder to locate when harm occurs.

And when these systems fail, which they inevitably will, civilians pay the price. We have seen it in Ukraine. We have seen it in Gaza. And we are seeing it again in Iran.

Meanwhile the technology continues to accelerate, propelled by defence budgets, venture capital and an increasingly close alliance between Silicon Valley and the military.

The speed of that shift should give democratic societies pause. Instead, the debate is only just beginning.

Subscribe now

✊ How to fight back

⏸️ What if the most powerful technology ever created… slowed down a bit? PauseAI is demanding politicians and companies pause on AI development until proper international safety agreements are in place.

👉 Join their network.

❌ You can still #CancelChatGPT after OpenAI’s partnership with the department of war. Here’s every reason why, from historian Ruter Bregman.

👉 Here’s every step how, from the aptly named campaign QuitGPT.

🔌 Right now, a tiny handful of American tech companies are shaping the future of AI. I don’t know about you, but I don’t trust ‘em.

Pull the Plug agrees. They bring together juries of ordinary people (yes, you) every Wednesday at 7pm on Zoom to discuss the future of AI and how citizens can take back control.

👉 Sign up here.

And they’re not just talking. They recently helped organise the largest AI protest march London has seen yet. Go on!

Thank you for reading.

See you next time,
Team Citizens

Share

About The Citizens Understand:
In an era where technology is reshaping democracy faster than laws can keep up, the Citizens Understand exists to cut through the noise and make complex systems legible. If there’s something you’d like to understand, email lillian@the-citizens.com.

For two decades, child sexual exploitation online has largely been framed as a moderation problem. But the renewed focus on the Epstein files has reopened a more uncomfortable question: what if this is not a failure of enforcement but a feature of the system?

The files exposed how closely the broligarchy - the billionaires who control the tech platforms shaping public life - intersected in Epstein’s orbit. As our founder Carole Cadwalladr wrote: “Epstein’s world is our world. That’s the darkest revelation of these files. He wasn’t an aberration. He was our culture made flesh.”

That culture - an obsessive, pervasive sexualisation of teenage girls, and to a lesser degree boys - does not exist outside the internet. It is threaded through it.

Our culture eroticises teenagers for profit. Our digital systems identify demand, connect users and amplify what keeps them engaged. Recommender algorithms prioritise what keeps users interacting, rather than assessing the nature of that interaction.

Generative AI has added a new layer of risk, laid bare when Musk’s Grok was used to generate non-consensual sexualised images of women and girls. Meanwhile, tech giant Meta faces trial over allegations it connects predators to children, and governments are scrambling to catch up. How in the hell did we get here?

In the third edition of the Citizens Understand, we zoom out to look at how a profitable market around sexualised youth intersects with engagement-driven platform design, and whether the current crackdown from government goes far enough.

The market and the machine

Engagement-driven design doesn’t distinguish between curiosity and harmful fixation. Platforms register what a user lingers on, searches for or clicks, then feed them more and more of it - nudging them towards increasingly extreme material. With “teen” among the most searched terms on mainstream pornography sites, a single search can quickly surface more explicit categories and connected account.

Research shows exposure can precede deliberate intent. A Finnish organisation, Protect Children, posted anonymous questionnaires on the dark web to reach people viewing illegal material across several countries, including the UK. Of the 4,549 individuals who admitted to viewing child sexual abuse material, more than half said they had not been actively seeking abusive images when they were first exposed to them.

At the same time, recorded offending continues to rise. In England and Wales, more than 850 men are arrested each month for online child abuse offences. The Internet Watch Foundation reported that 2025 was its worst year on record for online child sexual abuse material.

This is now playing out in court, as Meta currently faces a jury trial in New Mexico — its second major trial of 2026 over alleged harms to young users.

The state alleges the company knowingly enabled predators to connect with children, pointing to design choices that prioritised engagement over safety, as well as unmoderated groups linked to commercial sex and the buying and selling of child sexual abuse material.

The lawsuit follows a two-year Guardian investigation, published in 2023, that included more than 70 sources from survivors and their relatives to prosecutors to content moderators. It found repeated claims that Facebook and Instagram had become “major sales platforms for child trafficking”.

In the same year, a Wall Street Journal investigation found that Instagram’s algorithms actively promoted illicit content and connected accounts dedicated to the buying and selling of child sexual abuse material.

We scrutinise the unchecked power of Big Tech to power movements pushing back.

👉 Help us continue.

Recommendation to generation

Generative AI had already been used to create child sexual abuse material, but at the beginning of this year, the world caught up.

In January, it emerged that Grok - the chatbot developed by Musk’s company xAI and integrated into X - had rolled out a new image-editing feature. The tool was swiftly used at scale to generate sexualised images of women and children. Analysis conducted for the Guardian found that around 6,000 requests an hour were being made to alter images to show women in bikinis or revealing clothing.

X initially responded by restricting image-editing replies to paying subscribers. Following further backlash, Ofcom opened an investigation into whether UK laws had been breached. X later announced that Grok would no longer be able to edit photos of real people to appear in revealing clothing in jurisdictions where this is illegal.

While restricting the feature was welcomed, some campaigners argued the response did not go far enough. Journalist and campaigner Jess Davies described it as “pathetic”, telling BBC News: “They’re just trying to do as little as possible within the loose legal guidelines that there are.”

Is the government’s response enough?

This week, Keir Starmer called online misogyny a “national emergency” as he announced a crackdown on AI-generated abuse.

AI chatbots will be brought under the Online Safety Act, exposing companies to heavy fines or even blocking if they fail to prevent harm.

The government has also ordered that deepfake nudes and non-consensual intimate images must be removed from platforms within 48 hours of notification, with penalties of up to 10% of global revenue for companies that fail to act. The aim is to shift the burden away from victims, who currently have to repeatedly report the same image as it resurfaces.

Campaigners welcomed the move. Sophie Lennox of Everyone’s Invited told us: “Being a victim of deepfake abuse should be viewed as being just as traumatic as in-person abuse.

Lennox described the proposed law as “the start of building tech that protects, rather than tech that enables abuse”, and said the “relentless efforts” of survivors had been “crucial” in forcing the issue onto the national agenda.

At the same time, ministers are accelerating proposals that could enable a social media ban for under-16s, framing it as an immediate safeguard against exploitation and harmful content.

But some argue that restricting access risks treating the symptom rather than the cause. If access is removed, companies face less pressure to build safety by design.

Journalist and online safety campaigner Adele Walton told us: “A blanket ban on social media for under 16s risks subjecting young people to a cliff face of harmful content once they turn 16.

“What young people need is the government to actually hold big tech companies to account for the harm they’re perpetuating.”

She argues a long-term solution would be a ban on the addictive business model. “This, in practice, could eradicate the online harms - be it self-harm material, grooming and coercion, eating disorder content, or social media addiction more broadly - that people of all ages are facing right now,” Walton said.

“A wholesale social media ban only risks punishing children for our failure to regulate social media platforms over the past 20 years.”

✊ How to fight back

👉 Join International Justice Mission’s campaign calling for technology that protects children - and tell your MP that devices should be free from child abuse.
https://ijmuk.eaction.org.uk/protect-now

👉 Educate yourself with Ctrl Alt Reclaim’s report, featuring testimonies from youth activists across Europe calling for safety-by-design standards, fairer platforms and meaningful youth participation in digital policymaking.
https://ctrl-alt-reclaim.org/reports/young-voices-reshaping-the-digital-world

🤓 Learn more

📚 What is the human cost of our digital world? Read Adele Walton’s Logging Off, a sharp examination of how engagement-driven platforms profit from anxiety, insecurity and vulnerability.
Order it from Waterstones.

🎥 Is Regulation the Big (Tech) Fix? From the Citizens archive, join journalist Manasa Narayanan as she speaks to Marietje Schaake about whether existing tech laws are working and whether AI can be meaningfully regulated.
Watch here.

Thank you for reading.

See you next time,
Team Citizens

Share

About The Citizens Understand:
In an era where technology is reshaping democracy faster than laws can keep up and power is increasingly exercised through platforms, the Citizens Understand exists to cut through misinformation and make complex systems legible. If there’s something you’d like to understand, email lillian@the-citizens.com.

Welcome back to the Citizens Understand.

In our second edition, we trace how the Epstein files connect Peter Thiel, Palantir and Peter Mandelson - and how those links expose the tight weave between extreme wealth, political authority and foreign influence.

Note: Inclusion in the Epstein files does not in itself constitute evidence of criminal wrongdoing.

As fresh disclosures from the Epstein files ripple through British politics, media attention has focused on Peter Mandelson, now under investigation by the Metropolitan Police over alleged leaks of market-sensitive information.

The files shed light on a huge network of political, financial and corporate relationships, including - you guessed it - Peter Thiel, founder of Palantir, the firm that has its tentacles in practically every U.K. government department.

👀 At a glance

• Palantir is a US surveillance and defence analytics firm now holding UK government contracts worth hundreds of millions of pounds.

• Its co-founder and chair, Peter Thiel, appears repeatedly in the latest Epstein files and accepted tens of millions of dollars in Epstein-linked investments via a Thiel-controlled fund.

• Mandelson, through his lobbying firm Global Counsel, played a key role in Palantir’s UK positioning - including arranging Keir Starmer’s visit to Palantir’s Washington headquarters.

• The UK’s intelligence and parliamentary committees are now under pressure to examine conflicts of interest and national security risk.

⚙️ What is Palantir, again?

Palantir Technologies is one of the most powerful and controversial tech firms of the last two decades. Founded in 2003 and co-founded by billionaire investor Peter Thiel, it built its reputation supplying data-fusion software to US intelligence agencies and the military during the ‘war on terror’.

Palantir markets itself as an “operating system for government”. In practice, its tools have been used to enable drone warfare, embed policing systems later accused of fuelling institutional racism, and assist U.S. Immigration and Customs Enforcement, actively generating leads to locate and deport migrant families.

👉Read our full explainer on Palantir.

Deadly deportations in the US. Surveillance in Gaza. Palantir now deeply embedded in UK national infrastructure. We have questions.

✍️ Please sign and share our petition calling for a full Parliamentary debate on Palantir.

🗂️ What do the Epstein files show about Peter Thiel?

The latest tranche of records released by the US Department of Justice places Peter Thiel within Jeffrey Epstein’s network of wealthy and influential contacts.

Thiel’s name appears at least 2,200 times across emails, calendars and documents, with material showing the two men meeting and corresponding repeatedly between 2014 and 2017 - long after Epstein’s 2008 conviction for soliciting a minor.

Records also detail significant financial dealings. Epstein invested millions of dollars into funds managed by Valar Ventures, a firm co-founded by Thiel, becoming a limited partner and receiving what records describe as “confidential” and “time-sensitive” investment opportunities. Those dealings continued until shortly before Epstein’s final arrest in 2019.

Thiel has denied wrongdoing. But the files show a long-running business relationship that has only now come into view as Palantir’s role inside the British state has expanded.

💼 Where does Peter Mandelson fit?

Between 2009 and 2010, while serving as business secretary, Peter Mandelson is alleged to have shared confidential government information with Jeffrey Epstein - disclosures now under investigation by the Metropolitan Police. What has drawn less attention is what followed in the years after.

In 2018, Palantir hired Global Counsel, the lobbying firm Mandelson co-founded and still part-owns, to help position the company as a credible partner to the British government. Though Mandelson had stepped back from day-to-day management, he remained Global Counsel’s president and chair of its international advisory board.

At the time Palantir engaged Global Counsel, both Mandelson and the company’s co-founder Peter Thiel had documented links to Epstein. How those relationships were assessed, or whether they were assessed at all, remains unclear.

In February 2025, shortly after taking up his post as ambassador to Washington, Mandelson was present when Keir Starmer visited Palantir’s headquarters and met chief executive Alex Karp.

At the time, Palantir was still a Global Counsel client. Seven months later, Palantir received a £240m Ministry of Defence contract without competitive tender, under a defence and security exemption.

The Foreign Office says it holds no correspondence relating to Mandelson’s role in arranging it.

Become a paid subscriber and support journalism that scrutinises the growing alliance between state power and Big Tech.

Subscribe now

🇷🇺 How does this link to Russian interference?

“His connections with Russia raises questions because he is the perfect agent of influence. He can compromise people, he can blackmail people, he knows everybody, and he can move their money around.” - Rory Stewart, The Rest is Politics: How Epstein Poisoned our Politics

A growing body of evidence shows Jeffrey Epstein acted as a bridge between wealth, politics and foreign power, with proximity to figures connected to the Kremlin.

The files and intelligence-linked assessments have come to suggest Epstein advised the Kremlin on sidestepping Western sanctions, used his network to open doors between Russian-linked individuals and Silicon Valley companies and relied on Russian networks - later tied to money-laundering investigations - to process payments to trafficked women.

Seen in that light, it becomes clearer why the revelations around Peter Mandelson are treated as national security matters. And it inevitably affects how we understand Epstein’s financial relationship with Peter Thiel. While Epstein was investing in funds linked to Thiel and receiving preferential, confidential opportunities, he appeared to represent a route through which foreign influence could travel.

For us, this is the crux. If the Mandelson-Epstein affair shows anything, it is how urgently Britain still needs a proper reckoning with Russian interference in its politics, as long-running questions remain unanswered and companies linked through these networks hold sensitive state contracts.

We think the public deserves clarity. As MPs prepare to debate Russian interference today, the government could choose transparency and call a full independent inquiry. Here, we set out why that step remains necessary.

🤓 Further resources

What do Mandelson’s ties to Jefferey Epstein reveal about the influence of the rich and powerful in British politics? How do wealthy elites exploit political systems, and what structural reforms can protect democracies from corruption?

Join Rory and Alastair as they answer all these questions and more.

This is Jeff’s World. And we’re living in it.

For a raw and personal reflection on what the Epstein revelations say about power, impunity and the “paedophiliac culture” that enabled him, read our founder .

How to Survive the Broligarchy

We all live in Jeffrey Epstein's world

A note on who I am: I’m an investigative journalist who’s spent a decade reporting on the collision of technology and democracy including exposing the Facebook/Cambridge Analytica scandal for the Guardian and the New York Times. Two years ago, I called the alliance of Trump, Silicon Valley and a global axis of autocracy: a tech bro oligarchy, aka the Br…

Read more

3 months ago · 2035 likes · 328 comments · Carole Cadwalladr

✊ How to fight back

Peter Thiel’s Palantir is wiring itself into critical UK services - with over £630m in government contracts. If these tools are going to sit at the heart of the state, democratic oversight has to catch up.

👉 Join our call for a full Parliamentary debate on Palantir’s role.

Campaigners are also challenging Palantir’s position inside the NHS. The Good Law Project is organising to push back against the use of its technology in our public healthcare.

👉 Add your name and say no to Palantir in the NHS.

See you next time,
Team Citizens

Share

About The Citizens Understand:
In an era where technology is reshaping democracy faster than laws can keep up and power is increasingly exercised through platforms, The Citizens Understand exists to cut through misinformation and make complex systems legible. If there’s something you’d like to understand, email lillian@the-citizens.com.

In response to our poll, you told us you want clearer explainers on the biggest stories in the news - alongside practical ways to take action. So we’re launching a new newsletter focused on exactly that.

Welcome to the first edition of the Citizens Understand.

Every other week, we’ll cut through noisy tech coverage to break down one complex issue shaping technology and politics, and a curated selection of further reading to help you dig deeper.

Most importantly, we’ll show you how you can respond. Because understanding is only part of the story. This newsletter is about turning information into action - and finding ways to push back, even when Big Tech feels too big to take on.

This week, we’re starting with Palantir: a powerful US surveillance company now deeply embedded in UK defence, borders and public services.

Become a paid subscriber and support independent journalism challenging Big Tech and state power.

Get 50% off

Why a secretive US surveillance company is now embedded in Britain’s military and public services

Named after the all-seeing stone in Lord of the Rings, Palantir Technologies has become one of the most notorious technology companies of the last two decades and one of the most powerful.

Its systems are now deeply embedded in US immigration enforcement, military operations abroad, and, increasingly, the British state.

Now, without public debate or competitive tender, the UK government has handed one of the most sensitive parts of our national infrastructure, the Ministry of Defence, to a company whose political alliances are deeply entangled with the Trump administration.

Let’s break down how we got here 👇

👁️ What is Palantir?

Palantir was founded in 2003 and co-founded by billionaire investor Peter Thiel, a libertarian tech magnate, major Republican donor and long-time ally of Donald Trump.

Palantir’s UK operations are led by Louis Mosley. Mosley is the grandson of Oswald Mosley, who led the British Union of Fascists in the 1930s.

The company builds data-mining platforms capable of ingesting vast amounts of information, including GPS data, phone records, social media activity and biometric data.

Palantir has two main products:

• Foundry, used by corporations and public bodies to manage large-scale operations

• Gotham, marketed as an “operating system for global decision-making”, designed primarily for governments, police forces and militaries

Gotham’s function is deceptively simple: it takes fragmented datasets held across multiple agencies and fuses them into a single, searchable intelligence web.

That efficiency is precisely the appeal. It is also the danger.

Beyond storing data, Gotham connects information in ways that allow authorities to move beyond record-keeping and into continuous profiling - building detailed personal profiles, mapping social networks and tracking peoples’ movements.

⚔️ Why are we talking about Palantir now?

Because the UK Ministry of Defence (MoD) has just signed its largest ever defence data contract - a £240 million “strategic partnership” - with Palantir.

It was announced during Donald Trump’s state visit to the UK.

Under the deal, Palantir will provide “data analytics capabilities supporting critical strategic, tactical and live operational decision-making” across the British armed forces. The contract is three times larger than the previous MoD agreement signed in 2022, and cements Palantir’s role at the heart of UK military modernisation.

Despite Palantir’s deep embedding across government systems, there has been little meaningful public scrutiny of its role and risks.

✊ Sign our petition calling for a full Parliamentary debate on Palantir’s use in UK services.

⛓️ Immigration, enforcement and the “kill chain”

Palantir’s rise is inseparable from Donald Trump’s return to power.

Since Trump re-entered the White House, Palantir’s US government business has surged. In March, Trump signed an executive order requiring federal agencies to share data on Americans. Palantir was selected to help make that possible.

One of the biggest beneficiaries has been U.S. Immigration and Customs Enforcement (ICE), the federal agency responsible for immigration enforcement, detention and deportation inside the United States.

Over the past decade, ICE has spent more than $200 million on Palantir contracts, now using tools to map neighbourhoods, generate profiles on individuals and prioritise potential deportation targets.

The risks of this model were described starkly in our reporting last year, where we interviewed a former Palantir employee Juan Sebastián Pinto who explained that the company’s technology forms part of what he called an AI-assisted “kill chain”: systems designed to pull together data from across government and turn into operational decisions.

Those decisions translate into immigration raids: coordinated enforcement operations in which officers are dispatched to specific homes or neighbourhoods to detain people for deportation.

These raids use the firm’s app, ELITE, to locate and prioritise people for deportation, using addresses from medical records to do it.

Over in the UK, Palantir has been awarded a £330 million, seven-year contract to build the NHS Federated Data Platform, consolidating patient data from up to 240 trusts.

This raises an unsettling question: what happens when a company whose tools are used to mine health data for immigration enforcement abroad is entrusted with Britain’s most sensitive public dataset at home?

Our 50% discount is still live. Until 31 January, you can support the Citizens for £2.50 a month or just £30 for the year.

Upgrade to paid

🔍 Surveillance, targeting and Gaza

The same surveillance infrastructure used to track and target people inside the United States has also been deployed in Israel’s military campaign in Gaza.

In January 2024, the company signed a “strategic partnership” with Israel’s defence ministry, formalising a relationship that stretches back more than a decade.

Since at least 2013, Palantir has worked with Unit 8200, Israel’s elite military intelligence unit responsible for mass surveillance, codebreaking and cyber operations. That infrastructure forms part of the intelligence architecture that has enabled Israel’s genocide of Palestinians.

Palantir’s AI systems sit within a wider surveillance ecosystem used to identify and prioritise “targets”. Investigations by +972 Magazine revealed that tools such as Lavender assigned people in Gaza numerical scores estimating their likelihood of being affiliated with armed groups, based on sweeping indicators such as age, location or patterns of communication.

Analysis: The risk is not failure but permanence

Once embedded, systems like Palantir’s are difficult to unwind, reshaping expectations of speed and scale and making it politically costly to step back.

That risk is compounded when the software is proprietary and foreign. Embedding American surveillance infrastructure into Britain’s defence systems means the UK does not fully control the tools it relies on, at a moment when those tools are tied to political actors openly hostile to democratic norms.

When the same technology used in the US to track migrants and profile protesters is embedded in Britain’s defence systems and used to consolidate NHS data, the risk is not failure but permanence, where surveillance becomes normalised and accountability disappears.

🗞️ From the archive

A snapshot of our reporting on Palantir over the years documenting its growing influence across military, security and public institutions.

Revealed: Involvement of Palantir & Faculty in UK Public Sector

How the tentacles of Palantir continue to encircle the NHS

‘The Default Operating System’: Palantir’s £245 Million Expansion into the UK

Inside Palantir: Profits, Power & The Kill Machine

✊ How to fight back

Palantir’s expansion in the UK’s government infrastructure and public services has happened with almost no democratic scrutiny.

👉 Sign our petition calling on Parliament to hold a full public debate on its use, risks and accountability.

Share

See you next time,
Team Citizens

About The Citizens Understand:
In an era where technology is reshaping democracy faster than laws can keep up and power is increasingly exercised through platforms, The Citizens Understand exists to cut through misinformation and make complex systems legible. If there’s something you’d like to understand, email lillian@the-citizens.com.