Correct but Flock’s own system, if their public statements can be trusted, doesn’t do facial recognition itself. The main use is tracking vehicles using license plates and all kinds of visual cues, and Bluetooth/Wi-Fi devices by their MAC address and properties, and linking this huge swath of data together. Still, faces are recorded as video that can be checked manually or processed by third-party tools, including “AI” facial-recognition ones. Therefore, if a pedestrian destroys all Flock cameras and has Bluetooth earpods on them, there will be an easy way to link the events, and then the police will check if their face is visible in surviving footage and check the country-wide database if the person ever used the earphones in a vehicle near the cameras.
Yeah that’s weasel shit, “our systems don’t do facial recognition” can just mean the cameras don’t have facial recognition built in and that’s all done off-cam on the backend systems.
Facial recognition is still happening, but they can weasel out of admitting it by stipulating the camera isn’t capable of it.
Yes, the searchable DB is the big problem. But they work like regular IP cameras too: do you think police would buy them as much if they didn’t? Or did stream but blurred out anything that wasn’t a vehicle? If a crime happened recently near a Flock camera and they know when, they can absolutely use whatever it saw. Not to mention, third-party “AI” tools can be always used on raw footage from any camera for face recognition, event classification etc. so you’re not safe even as a pedestrian without the police having “probable cause”. And the camera going offline is probable cause*.
* Here, probable cause would ideally mean legally justifiable reasons to check the recording. However, in practice, it’s just any excuse for a police officer to justify spending time retrieving and watching the footage.
All the news stories I’ve seen talk about concerns over tracking movement. This is the first I’ve heard about people worried about the actual video.
Look up Benn Jordan’s videos on these things. They are NOT just license plate readers.
Correct but Flock’s own system, if their public statements can be trusted, doesn’t do facial recognition itself. The main use is tracking vehicles using license plates and all kinds of visual cues, and Bluetooth/Wi-Fi devices by their MAC address and properties, and linking this huge swath of data together. Still, faces are recorded as video that can be checked manually or processed by third-party tools, including “AI” facial-recognition ones. Therefore, if a pedestrian destroys all Flock cameras and has Bluetooth earpods on them, there will be an easy way to link the events, and then the police will check if their face is visible in surviving footage and check the country-wide database if the person ever used the earphones in a vehicle near the cameras.
They can’t.
Yeah that’s weasel shit, “our systems don’t do facial recognition” can just mean the cameras don’t have facial recognition built in and that’s all done off-cam on the backend systems.
Facial recognition is still happening, but they can weasel out of admitting it by stipulating the camera isn’t capable of it.
Yes, the searchable DB is the big problem. But they work like regular IP cameras too: do you think police would buy them as much if they didn’t? Or did stream but blurred out anything that wasn’t a vehicle? If a crime happened recently near a Flock camera and they know when, they can absolutely use whatever it saw. Not to mention, third-party “AI” tools can be always used on raw footage from any camera for face recognition, event classification etc. so you’re not safe even as a pedestrian without the police having “probable cause”. And the camera going offline is probable cause*.
* Here, probable cause would ideally mean legally justifiable reasons to check the recording. However, in practice, it’s just any excuse for a police officer to justify spending time retrieving and watching the footage.