The Secret Software Deciding Who Police Visit Next

Buried deep inside a nondescript office building in a city nobody thinks twice about, a small team of data scientists is quietly running software that decides who gets a knock on the door next. Here's the part that should stop you cold: they're not cops. They're not judges. They're not elected by anyone. And yet their algorithm is helping decide where patrol cars go, whose names pop up on watch lists, and which neighborhoods get flooded with blue lights at 2 a.m. It's called predictive intelligence targeting — and it is spreading across this country faster than anyone is willing to admit. **THE ALGORITHM KNOWS WHERE YOU LIVE** Here's how it works, and buckle up, because it's wilder than you think. These systems vacuum up staggering amounts of data — arrest records, 911 call logs, social media posts, even your shopping habits — and then spit out a heat map of who's "likely" to commit a crime. Sounds sci-fi, right? It's not. It's already happening in dozens of American cities. One major department reportedly flagged nearly 400 people on a secret list, many of whom had never been convicted of anything. Not arrested. Not charged. Just *flagged*. Their names sat in a database, waiting. Data scientists inside these companies will tell you it's math. It's neutral. It's objective. It is not neutral. Not even close. **GARBAGE IN, LIVES RUINED** Here's the dirty little secret the tech brochures don't mention: these algorithms learn from old arrest data. And old arrest data reflects decades of over-policing in Black and brown neighborhoods. Translation? Feed a biased system a biased past, and it predicts a biased future — then calls it science. It's bias with a receipt. Critics are screaming from the rooftops. Civil liberties groups have filed lawsuits. Researchers have published study after study showing these tools can *amplify* racial disparities instead of fixing them. One analysis found that a person's zip code mattered more than their actual behavior in some of these models. Read that again. Where you sleep could matter more than what you do. **THE TEAMS BEHIND THE CURTAIN** The predictive intelligence targeting teams themselves are often young, brilliant, and genuinely convinced they're saving lives. Many come from top universities with the best intentions. They build dashboards. They tweak models. They present findings to commanders who nod and deploy. But here's the gut-punch: nobody outside that room gets a vote. No public hearing. No transparency report. No way to appeal an algorithm's judgment of your character. Meanwhile, the companies selling this tech are raking in millions in taxpayer dollars — often with contracts that shield their methods from public scrutiny. Trade secrets, they say. Sorry, you can't see how we decide your fate. **WHAT'S COMING NEXT** And it's about to get worse. New AI tools are being pitched right now that don't just predict *where* crime might happen — they predict *who* might commit it, and even generate automated reports for officers before they've finished their shift. Some departments are already testing facial recognition layered on top. One mistake in the data, and the wrong person gets a target on their back. This isn't a movie. This is Tuesday. The question isn't whether predictive intelligence targeting works. The question is who gets to decide what "works" means — and whether the rest of us ever find out we were on the list. **OUR TAKE** Predictive policing was sold as a miracle cure for crime, but it's shaping up to be a surveillance machine with a math degree. Until these systems are transparent, audited, and accountable to the people they judge, every one of us is a data point waiting to be misread. That's not justice. That's a gamble — and we're all holding the losing ticket.