London police use Palantir AI to monitor their own officers

London's Metropolitan Police — Britain's largest law enforcement agency — has for the first time officially confirmed that it uses artificial intelligence tools from American company Palantir to monitor the behavior of its own employees. The system analyzes sick leave, absences, overtime, and other internal data to automatically identify officers with potential professional ethics problems. The Police Federation — the union representing rank-and-file police — immediately condemned the practice, calling it "automated suspicion."

Scotland Yard's admission resulted from a journalistic investigation by The Guardian. Until this moment, police leadership had consistently refused to confirm or deny any contacts with Palantir — a company whose reputation sparks sharp debate on both sides of the Atlantic. Palantir, founded with participation from Peter Thiel — one of Donald Trump's key allies in Silicon Valley — has long worked with American intelligence services and the military. Today, among its clients are the immigration service ICE, which conducts mass deportations under the Trump administration, and the Israeli army. For British human rights advocates, the mere fact of cooperation with such a company is already cause for concern.

But the essence of the problem runs deeper than simply choosing a contractor. For the first time, a major police service of a Western democracy openly applies algorithmic analysis not to suspects or citizens, but to its own officers. Formally, the goal is noble: Metropolitan Police is experiencing a deep trust crisis following a series of scandals — from the murder of Sarah Everard by officer Wayne Couzens to systematic allegations of racism and sexism documented in the damning 2023 report by Baroness Casey.

Police leadership is desperately seeking tools capable of identifying problematic employees before they commit serious crimes or violations. Palantir's algorithms, which analyze behavioral patterns — frequent sick leave, irregular overtime, unexplained absences — are intended to serve as such an early warning system.

However, this is precisely where technological efficiency collides with fundamental questions about workers' rights and the nature of suspicion. The Police Federation's choice of the phrase "automated suspicion" is not accidental. When an algorithm flags an officer based on statistical anomalies — say, too-frequent sick leave — this is not an investigation of a specific violation, but profiling based on data. An officer suffering from a chronic illness, or an employee going through a family crisis, risks being flagged by a system that does not understand context. Moreover, the very fact of total monitoring creates an atmosphere of distrust that may exacerbate the already low morale within London's police ranks.

There is also a legal dimension. British data protection law — UK GDPR and the Data Protection Act 2018 — establishes strict requirements for automated decision-making affecting people. If Palantir's algorithms generate recommendations affecting officers' careers or disciplinary procedures, this may require conducting a data protection impact assessment and ensuring employees' right to challenge automated decisions. It remains unclear whether all these procedures were followed, given that Scotland Yard concealed the fact of using the technology until the last moment.

The situation with Metropolitan Police fits into a broader trend that is rapidly gaining momentum worldwide. Law enforcement agencies from New York to Tokyo are increasingly deploying AI systems — for facial recognition, predictive policing, social media analysis. But the London case is unique because here algorithmic surveillance is directed inward at the organization itself. This creates a precedent that could spread far beyond the police: if it is permissible to algorithmically profile police officers, what prevents applying the same logic to doctors, teachers, civil servants?

For Palantir, this contract is another step in its expansion into the European public services market. The company already works with the UK National Health Service, which also provokes protests from human rights advocates. Each new contract strengthens Palantir's position as an indispensable provider of analytical infrastructure to the state — and simultaneously makes the question of how much democratic institutions should depend on a private American company with a very specific client list increasingly difficult.

London's experiment with AI surveillance of police officers poses an uncomfortable question to society for which there is no good answer yet. Can one fight institutional police problems using the same tools of total surveillance that society criticizes when directed at citizens? Or, to put it simply: who will watch those who watch the watchers?