Good. Scream as much as you want about facial recognition software; it doesn't give them any capabilities they didn't already have and if they do _anything_ with it it will eventually mess up and they will be eviscerated in court. What I'm more interested in is that all of these officers are now wearing cameras and sending an always-on video stream to home base. It's a trivial step to demand that that stream be saved.
> "it doesn't give them any capabilities they didn't already have"
Quantity has a quality all of it's own. While currently the police can attempt to observe and memorize every face they see during their patrols, and where/when they saw it, doing this without a technological assist will be fruitless. Police without technology can memorize the faces of a few select persons of interests, but for tracking the entire public they are relatively worthless. This technology threatens to change this.
> "if they do _anything_ with it it will eventually mess up and they will be eviscerated in court."
Why are you so confident of that? UAE police have been credibly accused of systematic human rights violations in the recent past, and I've seen no hint of eviscerating over that...
>> "... sending an always-on video stream to home base."
The trouble is, so long as the cops have the ability to switch off the camera/pull the battery/cover the lens, we lose all the benefits of wearable video for police.
Before beating up Rodney King 2.0, they just flip the switch. Then in court, "The officers' cameras had a technical malfunction at the time and there is no video of the incident."
But if they want evidence of you doing/saying something incriminating, you bet your boots those cameras are rolling.
This is actually an interesting systems design problem, wherein the users are adversarial but in total control of the hardware.
Not a good example. The Rodney King video played a large role in getting the officers acquitted.
Without the video, the prosecution would have had eyewitnesses and the injuries to King, which were pretty damning. The defense would not have really had anything to counter that.
With the video, the defense was able to take a step by step approach to the defense. They showed that the first blows were in response to King being uncooperative or aggressive, and were probably justified. They then framed the issue as when, if ever, did the beating switch from justified force against an aggressive suspect and turn to excessive force against a man who had stopped fighting and submitted?
Each time King was hit, he'd twitch or kick or flail an arm or a leg. The defense started with the first blows, which they were able to justify because King did start out aggressive, and then they went blow by blow, looking at the position of each officer and what he could see, showing that each saw one of those kicks or flails of King's, and that from his position the officer could not see that was an involuntary response to a prior blow. From what the officer could see, King was still violently resisting, and so that officer would take a swing. The defense would show that caused an involuntary response that made the next officer think King was still fighting.
They went all the way through the video that way, getting the jury to focus on the difficulty of pointing to any point and saying that this was where the line was crossed.
This worked, and they got acquittal.
No video, and the jury would have likely focused on the totality of what happened to King, and then it would be hard not to find excessive force.
I think the video also made the prosecution overconfident. I think they thought it was going to as simple as showing up, playing the self-evident video, proving that the accused were the officers in the video and that the video was real, and that would be it.
The only way to stop a bad guy with a camera is a good guy with a camera.
(And yes, the wording is intended to resemble a common talking point about guns. The difference is that a mistake made with a camera is often correctable, a mistake made with a camera is often final).
