Texas pauses state funding for Flock's AI camera network

Texas Governor Greg Abbott ordered state agencies to pause funding for Flock Safety's licence plate reader cameras, his office confirmed on Friday, as the Texas Tribune published an investigation into the state's role in building the network. The Tribune reported that the Motor Vehicle Crime Prevention Authority has directed at least $30 million in grants to local police departments, helping install at least 3,200 Flock cameras since 2023.

The cameras read more than plates. They use image recognition to record vehicle make, model and colour along with distinguishing features such as bumper stickers and dents — which is what turns a plate reader into a system that can follow a specific vehicle whose plate it was never given.

The order's reach is narrower than it sounds. Abbott's spokesperson said most funding for the cameras comes from the federal government, so a state-level freeze leaves the existing network in place and much of its future funding untouched. What it withdraws is the state's endorsement, at a moment when Texas cities and counties have begun cancelling contracts of their own.

The detail that should carry the most weight is a criminal case rather than a policy argument. A Lufkin police officer faces 100 counts of misusing public information for allegedly using the local Flock network to surveil eleven people. That is not a breach, a leak or a vendor failure. It is an authorised user doing exactly what the system was built to do, aimed at targets of their own choosing.

This is the characteristic failure of surveillance infrastructure, and the one least often designed against. Debate concentrates on whether the data should be collected and who may hold it. The recurring harm comes from people who already hold legitimate access, under weak logging, weak audit and weak consequences. A funding pause does not touch that. Query-level auditing, retention limits and independent review of access logs would — and none of those is what has been paused.