On the Creeping Normalisation of Palantir in AI Governance and Security Spaces

Stop Killer Robots is raising serious concerns about the growing presence of Palantir Technologies in international forums on global security, AI ethics and governance. Palantir is not a neutral technology company; it operates from the ideological positions of its founders and has publicly positioned itself as a champion of “hard power,” promoted the inevitability of “AI weapons,” and pushed for “a new era of deterrence built on AI.” AI weapons are inherently dehumanising and reduce people to numbers to be processed and then killed via algorithmic recommendation. Palantir is a company that has actively and deliberately chosen to embed itself in military and law enforcement operations that have caused serious harm to civilians. Accepting Palantir into norm-setting spaces is not inclusion, it is enabling creeping normalisation, and it carries real consequences.

Palantir’s record on human rights and civilian harm cannot be set aside. Human rights organisations, grassroots campaigns, and former Palantir employees have all repeatedly expressed concern regarding Palantir’s failure to respect human rights and international law, particularly through its contracts with military and law enforcement agencies that have been credibly linked to serious abuses.

In the United States, Palantir has contracted with the Department of Defense to deploy experimental AI technologies to generate targets for widespread US strikes on Iran, raising significant concerns around compliance with international humanitarian law. The company has licensed its surveillance and analytics products to the US Department of Homeland Security, including Immigration and Customs Enforcement (ICE), amid widespread reports of abuses in the current US government’s immigration enforcement campaign. Palantir also contracts with the Israeli Defense Forces for “war related missions”. In her 2024 report to the Human Rights Council, the UN Special Rapporteur on human rights in the Occupied Palestinian Territories found that aspects of Palantir’s partnership with the Israeli government were indicative of executive-level knowledge of unlawful use of force, and a failure to prevent it or withdraw involvement. Palantir has not stumbled into these contracts, it has sought them out.

Despite widespread concern about the ethical ramifications of the use of their technologies, Palantir has dismissed the debate around these technologies as “theatrical”. These are not the positions of a company that has reluctantly found itself caught up in difficult contexts, they reflect an explicit ideological stance that places Palantir in direct opposition to the goals of multilateral governance, international law, peace and security, disarmament, and civilian protection. This distinction matters when it comes to governance spaces.

We recognise that many technology companies operating in the defence and security space have links to potential harm, and that meaningful governance requires engaging with industry. Accountability, transparency, and binding regulation are essential tools for addressing this. But there is a difference between engaging companies that are navigating difficult contexts, or who hold problematic positions or contracts, and welcoming a company that has openly stated its opposition to the kind of multilateral, rule-based approach that international norm-setting efforts try to build. Giving Palantir a seat at the table does not strengthen the governance process. It sends a signal — to civilians harmed by their technologies, to human rights defenders, and to the broader international community — that the actors shaping the rules are the same actors profiting from the absence of them. 

We are calling on states, international institutions, and conference organisers to apply meaningful due diligence when involving industry in AI governance spaces to ensure that these efforts serve the purpose of building peace and protecting people, asking: Does this company’s conduct align with the values of multilateralism and civilian protection that underpin the UN’s mission? Has the company taken concrete steps to address credible concerns about its role in harm, and demonstrated a commitment to human rights?

In Palantir’s case, the answer is clearly no. 

It is up to all of us, civil society, states, and international institutions alike to hold the line against attempts to normalise inherently problematic companies and technologies, and to ensure that attempts to regulate AI are not shaped by those working against that very effort.