Policing's AI Boom: How Tech Giants Are Automating Law Enforcement—And the Risks

Thousands of vendors descended on Fort Worth this May for the International Association of Chiefs of Police Technology Conference to pitch the future of law enforcement to law enforcement itself. Facial-recognition cameras, automated license-plate readers, gunshot detection systems, AI-powered chatbots, report-writing tools, and algorithmic decision-making platforms all competed for departments' attention and budgets. The message was consistent: automation could free officers from tedious administrative work so they could focus on real policing.

But beneath that pitch lies a thornier reality. As police departments adopt increasingly sophisticated AI systems—from data collection to strategic deployment decisions—they're handing over judgment calls that have profound consequences for people's lives to technologies that lack meaningful oversight, transparency, or proven safeguards.

The Data Drowning Problem#

Police departments generate staggering volumes of information: body camera feeds, 911 dispatch logs, license-plate scans, CCTV footage, and case files. By 2019, the New York Police Department alone was accumulating roughly two years' worth of body camera footage every single week. Human analysts tasked with sifting through that deluge simply can't keep pace.

That's where real-time crime centers (RTCCs)—software platforms that aggregate data from multiple surveillance and monitoring systems—come in. Originally pioneered by the NYPD over two decades ago, RTCCs were once staffed by human analysts who manually compiled and summarized incoming data for patrol officers. Now, companies like ForceMetrics (whose co-founder is a former FBI cybercrime specialist), alongside giants like Axon and Motorola, are deploying AI to automatically extract patterns from oceans of data and recommend resource allocation.

"All the systems of record [used by police departments] are essentially antiquated," said Jason Truppi, co-founder of ForceMetrics, which markets its Velocity platform as an "AI-powered decision-assist" tool.

A Troubling Track Record#

The pitch isn't new. Police departments have experimented with data-driven systems for years—CompStat and PredPol chief among them—with the goal of removing human bias from law enforcement decisions. Instead, those systems ended up reinforcing and amplifying the very racial disparities and systemic biases they were supposed to correct.

Yet advocates for these newer AI platforms argue that earlier failures resulted from insufficient real-time data. Today, with vastly more information available, algorithms should finally deliver on the promise of unbiased decision-making.