Flock Safety: 71% License Plate Reading Errors

Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
Flock Safety Under Fire for Its AI Cameras
Flock Safety, a company specializing in surveillance systems, is facing a series of controversies. Recently, its cameras have been criticized for their inability to accurately read license plates, adding to other concerns regarding the capture of private conversations by its microphones. A recent report highlighted a major issue: Flock Safety's AI has a concerning failure rate.
The Roseville Police Department in California conducted a thorough investigation after receiving numerous complaints about vehicle reporting errors from Flock Safety's cameras. By examining approximately 1,500 reports, they discovered that 71% of license plate readings were incorrect. This numerical confusion, such as mistaking a 7 for a 1, led to serious errors, where innocent vehicles were reported as stolen.
Errors with Serious Consequences
These errors are not without consequences. One case in California illustrates the problem well: a citizen saw their car wrongly reported as stolen multiple times. Flock Safety's technology had misinterpreted a 9 as an 8, resulting in erroneous reports and inconveniences for the vehicle owner.
Flock Safety claims that its cameras are 96% accurate for vehicle counting and 92% for categorizing vehicle types. However, these figures do not pertain to license plate reading, for which the company does not provide detailed data. Requests for additional information have been made to Flock, but responses are still pending.
Potential Impact on City Adoption
The reliability of Flock Safety's cameras is being questioned, raising doubts about their utility for cities. The analysis by the Roseville Police Department was conducted between 2023 and 2024, a period during which AI models can evolve. While some models improve over time, others may degrade by ingesting incorrect data, particularly that generated by other AIs.
Flock Safety mentioned that the camera setup in Roseville was not representative of their usual use, but did not provide further details. Factors such as camera placement, weather conditions, and the state of license plates can influence the accuracy of readings.
A Challenge for the Future of Flock Safety
These issues are critical for any city considering adopting Flock's cameras. The company promises effective and accurate surveillance, but if errors persist in common situations, it could burden law enforcement rather than alleviate it.
Some cities, like Los Angeles, have already decided to terminate their contracts with Flock, sometimes opting for other providers. Public pressure plays a significant role in these decisions. If Flock's services continue to exhibit a high failure rate, it could prompt other police departments to reconsider their collaboration with the company, making it difficult for Flock to expand into new municipalities.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.