Quote:
Originally Posted by Varyon
In a society where the authorities do a lot of shady stuff but are nominally required to follow the law - or at least don't want the public to have irrefutable proof they're doing shady stuff, and/or want to conceal the identities of their agents - all security cameras could have hidden "redaction" backdoors so that they erase the presence of any "Men in Black" from their screens, making such agents invisible to the cameras (and to anyone monitoring them). Criminals with appropriate connections could make use of this redaction function to conceal themselves as well. You'd probably have frequent updates of the firmware as the protocols change once it is identified that the current method is live in the wild, of course. So, most criminals might have access to an older redaction exploit that only works on cameras that haven't been updated for a while, but the most well-connected would have 0-day exploits that work on nearly all cameras (but some government buildings would be beyond the bleeding-edge, or even have their cameras outright lack the redaction backdoors; in many cyberpunk settings, this would also apply to some corporations, but only for their most-secure locations).
|
While perhaps useful for a dystopian story environment to have a system that permits editing on the fly or even editing after the fact - chain of evidence being what it is, the first thing that an attorney is going to do is try to refute any video evidence as doctored if they can find any proof of this being the case. Such editing would leave traces.
Sadly, in the dawn of Deep Fakes, the better these fakes become, the more in question some of the videos may come to pass. If I recall, there have been a few movies produced (including an Arnold Schwarzenager movie if I'm remember correctly) where the "Hero" is an innocent framed by such technology. In a way, this becomes an instance of offense versus defense technological evolution. If ordinary video can be spoofed to where there is no way to determine fake from real, chances are good that video will not be permitted as the sole form of evidence.
But that doesn't deal with the original question. Video surveillance is not evidentiary - it is simply a way to automate surveillance to where you do not need eyes on the ground to catch sight of someone. If automated "Optical recognition database" programs permit one to search 10,000 faces looking for a list of faces - and FINDING someone who is on that list - that in effect acts as a force multiplier. A police force of say, 150,000 in a city of 8 Million people, but augmented by 2 million cameras - is a far cry from a city of 8 million with 150,000 police who do not have access to 2 million cameras and untold software apps who tirelessly search through a lot of video, isolate faces, compare them against a database, then flags the police officers that a "listed face" has been spotted.
Yes, there will be some false positives. Problem is, there likely will be more actual positives to make the use worth the while. If each "false positive has to be investigated by a live agent, the police no longer have to spend as much effort to spot those few faces out of a city of 8 million people. The cameras already did the grunt work.