Hackers breached a Flock Safety camera and exposed the system's facial recognition capabilities

Hackers gained unauthorized access to a Flock Safety camera, allowing them to analyze the internal mechanisms of the automatic license plate recognition system. The incident revealed technical details of data transmission and the architecture of device interaction with the company's cloud infrastructure.

During the hack, researchers found that the device uses standard communication protocols, which under certain conditions allow metadata to be intercepted. The Flock Safety system relies on computer vision algorithms to identify vehicles, their makes, models, and colors in real-time. The obtained data confirmed that the cameras transmit information to central servers through encrypted channels, however, vulnerabilities in the firmware can lead to a leak of device identifiers.

Analysis showed that the system processes thousands of requests per second, matching numbers with law enforcement databases. The device's internal logs contained information about network configuration and image processing parameters, which had not been publicly documented before. Experts note that such systems often have hidden debuggers that can be used to bypass protection.

Flock Safety has already begun a security audit after the report was published. The incident highlights the risks associated with the use of IoT devices in critical monitoring infrastructure, where recognition accuracy and data transmission speed are key parameters. Details of the discussion of the technical aspects of the hack are available in the Hacker News community.