Hacker News Reader

Comment by jsty | Hacker News Reader

As in many things, a blanket statement of 'not ready for use' isn't really the most helpful way to look at the situation. There's a range of ways such technology could be usefully deployed, even with current technological problems, that might aid law enforcement officers in the field. It's a very wide continuum from zero to "automatically locking people away for life based on a facial match".

As the article says, there are huge concerns about how such tools might be misused, whether legally or not. So the title might more accurately stated not as 'facial recognition software is not ready for law enforcement', but that 'law enforcement is not ready for facial recognition software'. I'm fairly sure there are countries out there where citizens do not lie awake at night in fear of their government, and where such technologies might be responsibly deployed as a tool in law enforcement.

View on HN

>> "I'm fairly sure there are countries out there where citizens do not lie awake at night in fear of their government, and where such technologies might be responsibly deployed as a tool in law enforcement."

This was true of many countries where it's no longer the case. Once a technology is available to a benevolent government, it remains available once that benevolence ends. It's wise to do what can be done to prevent and hinder misuse before it's too late.

Regardless of how you try to rephrase it, the fact remains the same: the technology doesn't work well enough.

This is going to sound horrible, but I suspect at least part of the reason people are unwilling to state that "facial recognition tech currently doesn't work well enough", is because there is a lot of money riding on it. Selling to law enforcement was probably a very large part of the plan for that sector. So whether it works or not, they kind of have to get law enforcement to use it.

It doesn’t sound horrible, but it is horrible.

Why does that sound horrible? That's how the world works.

I suppose it is horrible, but it's a very common situation.

...and law enforcement has shown it can't act responsibly with technology, especially when it comes to privacy.

You're right, but if "acting responsibly" was at all a prerequisite, they wouldn't even have their precious guns.

Or their cars.

Well enough for what?

You don't think there's _any_ conceivable way that law enforcement could utilize facial recognition tech without negative consequences? Not even, say, aiding manual review of security camera footage as part of an investigation into a preexisting suspect?

>You don't think there's _any_ conceivable way that law enforcement could utilize facial recognition tech without negative consequences?

Absolutely. Deploy the cameras around town and train it using pictures of their officers to catch misbehavior.

What do we do with all the data that includes everyone else’s faces?

If there is no potential officer misbehavior in the data, delete it after 60 minutes of recording.

"Regardless of how you try to rephrase it, the fact remains the same: the technology doesn't work well enough."

I don't suggest that anyone is looking to simply lock people up because of a 'face match' - rather flag an individual as 'possibly being someone' in which case this can be very useful information.

If for example they use it at border checks to 'flag' individuals, then a border agent can manually intervene, check the person's info etc..

So maybe that this not even possible to any degree of validity today, in which case 'the tech is not ready' ... but if the tech works to any measurable degree of reliability then it's possible to be put to use to reasonable effect - like anything (guns, tasers, DNA matching, fingerprints) - it has to be used properly.

Fingerprints seem to be a bonafide way to identify someone and can be used as evidence, it'd seem that facial recognition can't be used as evidence, but likely for other things.

If we could force the cops to always use the tool in a way that is consistent with its limitations, then you might be right.

Unfortunately, cops are not technologists. They're not mathematicians. They don't understand the dangers of applying a solution that is right 99% of the time to a population of 300 million people, and if each person averaged appearing on just one camera per day, that would mean three million false identifications per day.

How many real criminals are they really looking for on a daily basis, and which would warrant such a dragnet measure?

They can't handle a false positive rate like that. They can't be running around like chickens with their heads cut off, trying to verify the millions of bogus hits against the tiny fraction of potential valid ones.

It's the same problem that the TSA has, only these are real police with real guns. And real people are going to get killed because they get mis-identified, as opposed to just inconvenienced and pulled aside for extra scanning.

Do you really want to make the cops as bad as the TSA?

You don't have to 'force' cops if the proper parameters and processes are laid out.

Moreover, though cops can bend and break rules, they're not idiots.

For example, at border crossing, a 'fact match' with a name simply could flag someone for an interview, and a background check to see if it's a certain person.

An arrest has to be made on some kind of reasonable grounds if a 'face match' is not considered reasonable grounds, border guards (and cops for that matter) are quite aware of that and know they can't make the arrest on that basis alone.

Except ICE doesn't abide by your rules. You have no rights when being questioned or investigated by them.

At least, that's the way they operate.

Now, if you take them to court, maybe you can get a reversal of their action against you. But you have to be a citizen or legal resident to do that, and you still have to suffer the consequences in the meanwhile.

Yes of course I agree and we should be wary; it's bonkers that the US does not have clear and definitive rules for how the rules/constitution applies to non-residents etc..

Doesn’t some kind of facial recognition work well enough for Vegas casinos?

Instead of police being able to scan one face at a time as they walk their beats looking for suspects, an automated system could scan everyone in view and then alert that an suspect has been identified with xx% accuracy.

Every week on next door someone posts a video still of a vagrant or thief etc. either from a broken car window, hit and run, lurker, etc. Automation would help if the perps. We, as a society, might want to adjust the dials on punishment, given the efficiency, but we shouldn’t give up the chance to minimize these crimes when reasonable.

It is a bit more complex. There are few scientific studies on this topic. See this one: http://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a...

"We find that these datasets are overwhelmingly composed of lighter-skinned subjects (79.6% for IJB-A and 86.2% for Adience) and introduce a new facial analysis dataset which is balanced by gender and skin type. We evaluate 3 commercial gender classification systems using our dataset and show that darker-skinned females are the most misclassified group (with error rates of up to 34.7%). The maximum error rate for lighter-skinned males is 0.8%.

That’s pretty interesting. Given those stats, why can’t it be selectively deployed on light skinned perps till the system is trained better on other skin color classifications in order to achieve similar accuracies? And in the mean time catch some baddies.

Interesting, so given your implementation only targeting light skinned "suspects" would using this be racist?

I don’t think so. It would target more skin tones as it became better at identifying suspects accurately. The aim is to reduce crime regardless who commits it. We have a system which can help on a subset but not another. Why should we ignore the potential only because we cannot deploy it against everyone?

It’s like saying, well, since we can’t successfully prosecute big bankers, well also ignore smaller fry bankers who are sloppier cuz it’s not fair to them.

In the end everyone but the criminals benefit. Eventually one hopes the system is well enough trained so it can be deployed for all skin tones.

You're incredibly optimistic about how this will be used.

I have zero doubt that the second such a thing is deployed, they are going to be cataloging everyone's faces permanently then moving to track people next.

Given the lack of restraint common in policing (at least in the US), I don't think we want to start giving police probabilistic estimates that will direct them to categorize people as criminals. Even to look at them more closely, given that fatal SWAT raids are caused by prank phone calls. Imagine the impact of "well, the system told me he was 94% likely to be a criminal" turning into another excuse for these awful raids.

The false positive rate on fingerprints and the lack of understanding in DNA statistics further indicate that the enforcement/legal system has a very poor understanding of even basic statistics.

We also rely on witnesses with dubious memories and motives but people will never be held to the same standards as machines.

Reply on news.ycombinator.com