Flock’s cameras are the midterm fight. The prompt stack is the product.
Garrett Langley is touring “compromise” while Wired reports a police AI that turns neighborhood patterns into names and home addresses.
Flock Safety’s CEO wants the country to meet in the middle on privacy and safety while the company’s next product keeps expanding under him.
Garrett Langley told Fox News that talking about privacy or safety alone is “prioritizing the wrong thing,” and that America has to prioritize “compromise.” TechCrunch filed that as the latest stop on a media tour while cameras, drones, and license-plate readers draw a bipartisan pile-on. The Washington Post has identified 46 cases where officers were accused of using Flock technology for unauthorized purposes, including stalking partners. Langley apologized on CBS after hearing one alleged victim and still argued Flock did not create police abuse, only surfaced it and built tools to find it.
Hardware on poles, retention settings, municipal contracts, and candidates cutting ads are the public scrap. Axios puts Flock cameras beside data centers as a 2026 midterm boogeyman. Michigan Senate nominee Abdul El-Sayed tied opponent Mike Rogers to “mass proliferation of Flock cameras.” Bernie Sanders posted “STOP AI MASS SURVEILLANCE. STOP FLOCK.” Three House Republicans, led by Tim Burchett, introduced a bill to bar federal purchases of automated surveillance that uses facial recognition, biometrics, or plate readers, “including a Flock Safety camera.” House Democrats in Ohio, Tennessee, and Michigan are already running the same scrap. Municipalities have severed ties. Some people have sawed cameras off or painted the lenses. Flock told Axios it has been listening to customers and civil liberties groups, and it cites roughly 1 million investigations aided and around 10,000 missing people found. Default retention dropped from 30 days to seven, with a case code required to search. The ACLU notes Evidence Mode can still override the shorter clock, so the reform is only as real as that toggle.
A camera on a pole is something a voter can point at, and a retention slider is something a city council can argue about. That is why the fight looks like data centers: local, visible, and easy to put in a 30-second spot.
Wired reported, and Slashdot amplified, that Flock has built a prompt-based AI tool for police that can identify drivers and track vehicles by patterns of movement alone. For years the company told the public its technology “cannot recognize, identify, or track individuals.” The new system, Wired found, does both. Because it also reaches police case files, 911 dispatch logs, and commercial identity records, plates become names, home addresses, and relatives. It can search an area drawn on a map from a physical description. Wired describes 45 tools at the AI’s disposal: plate scans, camera metadata, arrest records, case files, dispatch logs, ballistics results, and commercial databases holding Social Security numbers, dates of birth, phones, emails, relatives, and associates.
Flock says the product is in testing with a small group of law enforcement partners and still in development, with capabilities that may not match what eventually ships. Flock’s testing hedge does not change the design.
One preloaded prompt asks for vehicles most seen in a neighborhood over the last 14 days in a daily timeframe, excluding whitelisted vehicles. Fill in the blanks. The output is a list of plates. Other tools convert those plates into names and home addresses. Another prompt asks the system to list everyone arrested more than twice in two years for “any offense,” exempting only narcotics, then map where they live, pull calls for service at those homes, and “do a workup on the top three individuals.” A “workup,” in Flock’s terminology, is a one-command background check that starts with a name and date of birth and returns vehicles, prior suspect listings, relatives, phones, and online accounts. A George Washington University law professor told Wired it is clear Flock has aspirations “far beyond ALPRs to become a digital platform for policing.”
An officer no longer needs a plate, a name, or a crime to begin. They supply a place, a stretch of time, and a pattern of behavior. The system is designed to hand back the people who fit.
Langley asks state regulators to criminalize illegal use of Flock data and says there is too often no regulation and no accountability. He is right that misuse exists. The Post’s 46 cases are the receipt. He is also selling a stack that lowers the skill required to turn a blank map into dossiers. Softening default retention while shipping a prompt that ends in three workups is two products sharing a brand, not the same strategy told twice.
Axios notes the camera backlash converging with data-center anger until AI becomes one of the most potent midterm issues. A Minnesota House candidate told Axios voters already burn about data centers, and linking the two “really, really burns people.” It still undersells the product change. Voters can hate a camera. Fewer of them will see a prompt that starts with everyone in an area who has an arrest record and ends with three dossiers chosen by software.
The midterm fight is the visible layer because that is what can be filmed and banned by name, while the prompt stack is how Flock becomes infrastructure for policing rather than a network of plate readers. Langley’s compromise language aims at the layer voters can see, while the Wired tool aims at the layer officers type into.
Compromise is the press tour. The product already has a blank for the neighborhood and a button for the workup.
Pingback: Hugging Face plumbing, Abbott, Flock | Bitter Fool