London’s Metropolitan Police turned a secretive American data firm’s artificial intelligence loose on its own ranks last month. In just one week, the tool sifted through internal records on attendance, expenses, building access, and complaints. Results stunned even insiders. Hundreds of officers now face probes for everything from shift-rigging scams to suspected rape.
Corruption topped the list. The software pinpointed 98 officers abusing the IT system that schedules shifts, twisting it for personal or financial gain. Another 500 got prevention notices on the same issue, according to numbers the Met disclosed. Forty-two senior leaders—from chief inspector to chief superintendent—drew scrutiny for dodging hybrid work rules. Twelve more failed to declare Freemason ties, a potential conflict in policing.
Three arrests followed. Two officers received gross misconduct notices. Assessments and investigations continue for the rest. The pilot drew from data the force already held, no new surveillance needed. Yet it exposed patterns humans might miss. Faster than legacy systems, as one report put it.
Palantir Technologies built the software. The U.S. firm, known for military and immigration contracts, has stirred outrage before. Critics decry its tools as budding surveillance machines. In Britain, a National Health Service data deal sparked backlash, though officials insist Palantir can’t sell or train models on that information. CEO Alex Karp dismissed overreach fears in February, noting U.S. government revenue soared 66% to $570 million in late 2025. Safeguards exist, he argued.
The Guardian broke the story on April 25, detailing the Met’s findings (The Guardian). Digital Trends followed up hours later, highlighting arrests and broader implications (Digital Trends). The Times called it an AI search rooting out rogues, with abuse of authority for sexual purposes among flags (The Times).
Pushback came swift. The Police Federation, representing rank-and-file, labeled it ‘automated suspicion.’ Officers shouldn’t face opaque tools that mistake workload or sickness for crime, they said. As of today, the group weighs legal action against the Met over the monitoring. Yahoo News reported the development just hours ago (Yahoo News UK).
This isn’t Palantir’s first police dance in London. Back in February, the Met confirmed using its AI on sickness, absences, and overtime to spot standards slips (The Guardian). That pilot aimed to lift culture. Now, with results in hand, expansion talks bubble. LBC noted the Met eyes broader rollout post-crackdown.
But questions linger. How accurate? Biases baked in? The software crunches patterns from existing data—roster logs, badge swipes, complaints. No faces scanned, no body cams parsed. Still, black-box risks loom. Officers challenge flags in human reviews. Three arrests suggest solid hits. Yet false positives could erode trust.
Broader trends amplify the stakes. Britain’s Financial Conduct Authority deploys Palantir against financial crime, defending the contract in March before lawmakers (Reuters). Policing Insight echoed the Met’s haul: work-from-home cheats to corruption (Policing Insight). Cybernews warned of misconduct risks flagged across policy breaches to rape allegations (Cybernews).
Overseas, AI hunts cop misconduct too. A 2022 Northwestern study used machine learning on historical cases like Chicago’s Ronald Watts crew to spot bad-apple networks (Forensic Magazine). U.S. departments tap AI for body-cam reviews, flagging force or language slips—Truleo processes footage daily (Fox News). But pitfalls persist. A 2020 Wall Street Journal piece flagged early intervention systems predicting officer trouble, yet biases shadowed them.
Palantir’s Gotham platform aggregates vast civilian data for U.S. cops—names pull addresses, associates, risk scores. X posts buzz with fears of ‘guilty until proven innocent’ (X post by Jason Bassler). In Georgia, automated bus-lane fines curb corruption cleanly, one user noted.
So where next? Met brass praise the speed. Federation cries foul. Pilots prove potent. But scaling demands transparency. Officers demand explainable AI, auditable flags. Without it, resentment festers. Public trust hangs in balance—cops policing cops via code. Irony bites.
Three arrests. Hundreds probed. One week. Technology just redrew the thin blue line.
WebProNews is an iEntry Publication