Flock’s AI Cameras May Have Misread Over 70% of License Plates, Creating New Headaches for Cities
New data suggests Flock surveillance cameras have an abyssal recognition rate. Here’s what that’s a major problem for drivers.
The police department of Roseville, California, conducted a study on Flock Safety's surveillance cameras, finding a shocking 71% failure rate. The cameras often confused one license plate number for another, leading to incorrect identification of vehicles. For instance, a citizen reported that his car was flagged as stolen multiple times because the AI technology mistook a 9 for an 8.
This is a far cry from Flock's own claims of 96% vehicle count accuracy and 92% vehicle type categorization. The company has not provided more details on license plate recognition accuracy. While AI models can improve with better training, they can also become less accurate when exposed to bad data. Factors like camera placement, weather conditions, and license plate coverings can affect accuracy.
Cities like Los Angeles have canceled Flock contracts due to public pressure, citing the high error rate. This raises concerns for police departments about the reliability of Flock cameras, potentially complicating their adoption and sales to new cities.
Written by urgent.news from CNET's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
