There is a growing temptation inside Whitehall to treat Palantir Technologies Inc. not as a contractor, but almost as an unstoppable force. Something that is inevitable. So, faced with the real challenge of modernising defence and policing (and public administration) in an age of data weaponisation and artificial intelligence, it seems that successive ministers across all political spectrums have come to one conclusion. That the US-firm Palantir is simply too big, too knowledgeable and too far ahead of domestic British alternatives to resist.
This conclusion might be slightly understandable. But it carries deep risks and it carries clear moral concerns.
This month, The Citizens, an advocacy group, called for Parliament to hold a full debate on Palantir’s expanding role across UK defence, security, law enforcement and public services.
This is because, despite Palantir’s increasingly deep integration into Britain’s core state functions, there has of yet been no comprehensive parliamentary scrutiny of the costs, the data governance, the ethical risks or the long-term national security consequences of such a relationship. It is as if public accountability has failed to keep pace with technological opportunities and contracting-out.
This might be understandable. Palantir sells itself as invaluable. Via its software, it promises to fuse intel, data and operational evidence and make militaries unstoppable. They market themselves as battle-tested, interoperable with allies and proven in conflicts from Afghanistan to Ukraine. For a Ministry of Defence long criticised for fragmented data and slow decision-making, Palantir looks a sure be to be ahead of the rest in the war to ultra-fighting modernity. But nothing easily gained does not come without risk.
The most obvious danger is dependency. When contracts worth hundreds of millions of pounds are awarded without open competition, the state is not merely buying software. It risks handing away a core function of national power: how our information is organised, prioritised and acted upon. In contemporary conflict, data becomes the nervous system of military decision-making. Allowing a single foreign firm to design the way our system works risks concentrating strategic leverage in ways that should trouble anyone interested in such a democracy.
Ministers insist that data sovereignty is protected and that all information remains under UK control. Legal ownership, however, is not the same as practical autonomy. Once a proprietary system becomes the backbone of operations, switching away becomes costly operationally risky. The deeper Palantir embeds itself, the harder it becomes for government to renegotiate terms or impose meaningful conditions.
There is also a democratic deficit. Many of Palantir’s most significant UK contracts have been justified on grounds of urgency or national security, limiting public disclosure and parliamentary challenge. As a result, elected representatives are asked to take on trust assurances about how algorithms shape military priorities, policing practices or public-service decisions. That should be unacceptable. If proprietary systems increasingly influence targeting, threat assessment or resource allocation, democratic oversight must evolve accordingly.
This is not mere debate. What happens if Palantir militarily works with a country that becomes an enemy to the UK? After all, France’s deep defence relationship with Argentina before the 1980s Falklands war, particularly the sale of Exocet missiles and associated technical support, directly undermined British forces when those weapons were used against UK ships, killing 32 sailors and exposing a critical vulnerability in Britain’s campaign. Although Paris formally backed London once fighting began, the continued presence of French technicians in Argentina revealed how arms intelligence and embedded expertise can entangle allies on opposing sides of a conflict, blurring loyalties and weakening alliance cohesion at the moment it mattered most.
Ethical concerns also cannot be waved away. Palantir’s tools are explicitly designed to accelerate decision-making, including in military planning and targeting. Officials emphasise that humans remain in the loop. Yet experience from Israel across automated tareting systems shows that, under pressure, human judgement often drifts towards deference. The faster and more authoritative a system appears – as with the infamous Lavender system – the greater the temptation to follow the algorithm’s recommendations.
And, in the context of armed conflict or coercive policing, that erosion matters.
Palantir’s international record deepens these concerns. The company has been closely associated with US intelligence agencies and controversial immigration enforcement practices, and has faced sustained criticism from civil-liberties groups. It has also been criticised for its links to military operations in Gaza. Palantir denies any allegations of wrongdoing – and AOAV is not stating there are any – but if Britain hard-wires its security infrastructure to a firm so closely aligned with another state’s security apparatus, it imports not just software, but assumptions about power and the legitimacy of that state’s violence.
As such, according to the Financial Times, Palantir has emerged as a significant beneficiary of US President Donald Trump’s controversial immigration crackdown, securing at least $81m in contracts from US Immigration and Customs Enforcement since January 2025 as federal spending on enforcement has surged. The data analytics firm, which has worked with ICE for more than a decade, has won new deals directly linked to the administration’s harder line, including a $30m contract awarded in April to build an operating system for so-called “self-deportation tracking” and additional tools designed to streamline the identification, selection and apprehension of undocumented migrants.
This is why a parliamentary debate matters. It is not about technological nostalgia or knee-jerk suspicion of innovation.
It is – at its heart – about moral governance. Democracies do not defend themselves by adopting powerful tools blindly. Democracies require subjecting them to scrutiny proportionate to the power they confer.
In this light, Britain’s embrace of Palantir has run far ahead of public consent.
Action on Armed Violence has long argued that transparency, accountability and civilian protection must be central to security policy. The call from The Citizens is therefore deeply necessary.
Parliament should debate Palantir’s role in full view of the public, examining costs, dependencies, ethical safeguards and long-term consequences. The question is not whether Britain should use data-driven systems. It is who controls them, under what conditions and in whose ultimate interests.
Did you find this story interesting? Please support AOAV's work and donate.
Donate