1

There has been a lot of news lately about facial recognition use by police forces being banned or that it should be banned.

The obvious example is that a police officer who is motivated by personal reasons to misuse the capability could do so. While I agree there is potential for misuse like this... I would assume measures could be put in place to prevent this from becoming a regular occurrence.

So assuming that we can prevent that type of misuse what other reason could there be for banning facial recognition system from being used by police?

  • I believe false negatives would be the one issue. Unless the false negative ratio is way bellow the criminal-to-innocent ratio, facial recognition will "capture" a lot of innocent people. This can lead to an innocent being tagged as present on a crime scene even if he is miles away and some look-alike was there instead. – ThoriumBR Sep 18 at 14:16
  • I mean, I would assume some other corroborating evidence would be required to convict a person of a crime besides just facial recognition data. Physical evidence for example. – d4v3y0rk Sep 18 at 14:22
  • @MeckMK1 answer adds something interesting: data correlation can also be wrong, so a lot of people can be implicated because of multiple errors. Search Google for people that ended-up on the "Non-Fly List" because of typing errors... Imagine being labelled as terrorist because the Facial Recognition Software made an error and it agrees with the error on another system. – ThoriumBR Sep 18 at 14:26
  • There's even a case of Sebastian Khan, a 3 year old Canadian that is on that list by mistake because of his name. Typing a name wrong is less probable than your face looking like another face. Any human can tell if 2 names are not the same, but we fail to tell one person from another, so imagine a computer telling that two faces are not the same person. – ThoriumBR Sep 18 at 14:31
  • so it sounds like you are both saying though that is it not that they are using it... is is how they are using it that could lead to problems. so if we follow that logic it is not really bad for them to use it and so then instead of banning it by law we need to make laws on how it can be used... would you agree? – d4v3y0rk Sep 18 at 14:37
5

There are several comments on here regarding false positives (which are a valid concern) and the question rules out a rouge officer mis-using the technology for personal reasons, but even if the technology is perfect there are other avenues of abuse.

Perfect facial recognition for example could be used to identify participants of political rallies and used as means to target them later when they are isolated in order to suppress views people in power don't like.

It may be that you trust the authorities in your country not to abuse the system like this, but what about when the next government or the one that's in power in 10 years time orders the police to behave in this way? Once the genie is out the bottle it's difficult to put it back in.

  • The issue I have with this potential misuse is that none of the laws that are being enacted prevents private citizens or businesses from using this technology to do exactly what you are saying the government might use it for... so preventing the government from using it to fight crime is not going to stop political groups or businesses from using it to identify individuals that they can target for whatever their motives might be. – d4v3y0rk Sep 18 at 15:15
  • npr.org/2019/05/14/723193785/… "The ban does not affect personal, business or federal government use of facial recognition technology." – d4v3y0rk Sep 18 at 15:15
  • I agree private use may also be a cause for concern but since the question asks about the police that's what my answer addressed. – JustAnotherDev Sep 18 at 15:35
2

Who guarantees that facial recognition works fine? What would happen on a false positive?

Imagine the provider of the facial recognition software would guarantee that their system is 99.99% accurate. That sounds decent enough, right?

For the entire US population, this would mean roughly 32.000 people who would be accused wrongly. Of course, real-life face recognition software is not nearly as accurate, and may have much more problems.

For example, if a system only receives training data for white people, you may end up with strange results for other ethnic groups. Stereotypes like "all people from this ethnicity look the same" can then very quickly become a very real problem.

Aside from this, there are obvious privacy problems. Why is the police allowed to identify me? What if the information about my identity is cross-referenced with other data? What if that data is incorrect too?

The problem is the same as above: We know that technology is flawed, the police may know that technology is flawed, but the officer in the field doesn't consider that this one specific case may be a false positive, because the thousands of matches before may not have been.

2

I believe false negatives are a major concern. Even with 0.1% false negative ratio, it means hundreds of thousands of people would be miscategorized as suspect on the US alone.

Facial Recognition (and all AI tools usually sold together) are far, far from perfect. Coupled with data correlation (that can be flawed too) and incorrect understanding of the tools, you can end up flagging an innocent guy as terrorist, and ruining his life.

Even if this mislabelling does not ends in jail time (or worse), the inconveniences of being stopped at every security checkpoint is bad enough: it wastes the citizen and police officers time, wastes resources, and creates a defense to a wrongdoing guy when he is caught: he just needs to point to all false negative cases and say he is an innocent too.

  • I think a little wasted time is a small price to pay for increased safety but I understand what you are saying. I think if a technology like this were implemented everywhere and the incidents where it was inaccurate were used to continue to train it. the resultant system would get better over time. the training data is usually the source of bias in AI systems... correct? – d4v3y0rk Sep 18 at 15:25
  • It's not a little wasted time, it can take 4 or more hours of interrogation, plus missed connections, plus disruption of travel for anyone travelling with you. Imagine being innocent, detained for 6 hours, with no communication, and all your family waiting you to arrive, left without information from the airline or police. It's not a small price. – ThoriumBR Sep 18 at 16:48
  • As I read on an article today, once people spent months preparing data, cleaning mistakes, carefully labeling all records, and then feed to IA. Now people just throw on the algorithm every piece of data they can lay they hands on and hope for the best, as the sheer amount of data would get rid of issues by volume. It's not how it happens... – ThoriumBR Sep 18 at 16:50
0

X Y challenge:

It's not that they'll misuse it, it's that people feel that they have an innate right to privacy and anonymity that public facial recognition violates. People are worried that their every second of their day is going to become actively tracked and logged. It is the same argument for any privacy. It is not because they have 'something to hide' or are worried about consequences or police abuse... they just don't want the police to be allowed to essentially track every single person regardless of having a warrant or probable cause. There is a reason that you need a warrant to establish long term surveillance of someone, even if that surveillance doesn't require any specific action that violates that persons' rights.

https://www.newsweek.com/can-government-track-individuals-without-warrant-730388 This is a source that touches upon a closely related case, and includes Supreme Court reasoning against warrant-less tracking. Mass facial recognition is merely another iteration of using technology to facilitate police surveillance, expect more egregious. The police don't have to stick a GPS onto my car to track me, they can just decide that I'm worth looking into and go see where I was a week ago.

  • can the police take a picture of you without a warrant? can they use that picture in a lineup without a warrant? seems like this is just automating that capability and I don't think they need warrants to take pictures of people in public places.... – d4v3y0rk Sep 18 at 15:12
  • does the reasonable expectation of privacy prevent a cop for example from using his eyes to recognize you in a public place? I don't think there is really much difference. facial recognition in a case like that seems like a force multiplier. I.E. if we had enough cops roaming around with a picture of the person we are looking for. lets say 1 cop for every 5 people distributed evenly throughout a city. then there is every reason to believe they would find the person they are looking for. – d4v3y0rk Sep 18 at 15:32
  • Sure, but none of that is what you asked about. if you'd asked 'Why is there opposition to having cops walk around with photos of perpetrators?' the answer would be completely different, lol. You're arguing for something that I'm not arguing against here.There is a material difference between cops walking around with photos of a suspicious person and a permanent database categorizing everyone's facial biometrics that can be searched at their leisure. – Adonalsium Sep 19 at 12:40

Not the answer you're looking for? Browse other questions tagged or ask your own question.