Live data from Hacker News

Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

nytimes.com

131–140 of 256 posts

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#131

Earlier quoted context omitted.

> It is not inappropriate for police to be able to determine who they are interacting with if that person is a known criminal. Everyone is a criminal. You commit multiple crimes a day if you're like most people. There absolutely should be a clear delineation between what is possible and what is allowed when considering state surveillance. The 4th amendment: "The right of the people to be secure in their persons, hous…

Reasonable people can disagree on the definition of "unreasonable," however. Is it "unreasonable" for the Amber Alert system to co-opt the eyes and attention of ten thousand citizens around a kidnapped child? Hard to say. Most people seem to think 'no.' If that system could co-opt camera feeds instead of human attention, we could get most (all?) of the benefits of the existing Amber Alert without having to burn civil…

>If that system could co-opt camera feeds instead of human attention, we could get most (all?) of the benefits of the existing Amber Alert without having to burn civilian resources on it. Wouldn't that be a strict improvement over the current system?

In my area amber alerts are usually along the lines of "the non custodial parent has the child when he shouldn't". Reducing the resource cost of (and therefore threshold to utilize) that system or similar systems is not something I can get behind.

Finding and bothering specific people should be something that scales poorly enough that all the government organizations that engage in it are forced to pick and choose who they do it to. A low threshold is an invitation for abuse.

Edit: and I'm wrong because why?

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#132

This is a ridiculous article. Rekognition has been available for over a year, and I've been using it for that whole year. I've even integrated it into my own home security camera. Amazon is selling it to law enforcement, but it's not like LE get special service that no one else does. Amazon built a generic facial recognition system, and it works pretty well. Even if they weren't pushing towards the police and governm…

I'm not sure I follow because your home security system, correct me, is only near your home, not connected to millions of people collectively that are mechanical turked unwillingly. Them apples aren't oranges.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#133

Earlier quoted context omitted.

Only if you consider privacy as a means to an end rather than a goal by itself. Facial recognition has a negative effect on literally everyone's privacy.

Privacy is a means to an end. That end is peace-of-mind. People who have been damaged by random acts of crime may very well find more peace-of-mind sacrificing a bit of privacy to not have those acts repeated on their persons.

Should we install security cameras in public toilets? It would certainly reduce crimes such as assaults that occur in public toilets. There would be very little impact on our personal security.

Privacy is an fundamental aspect of our quality of life. Privacy is not of secondary importance to security but exists alongside it. If that were not so, then there are plenty of places we could be putting surveillance cameras in order to improve our security.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#134

I find this article a bit funny, and see it mostly as hype. I worked very revently on a large internal face detection service for a legal and compliance application at a large US company. Being pragmatic engineers, the first thing we did was to pilot test Rekognition against an in-house prototype built by modifying some open source deep learning approaches. Rekognition performed so poorly compared to our prototype (w…

interesting data point on its performance but the reason decision makers might be cloud happy is amazon has the resources to scale this to thousands of RPS and develop the backend codebase up to hundreds of engineer contributions for many years, even as key talent comes and goes. The same might not be true of your prototype especially if the core competency of your business is sales or legal compliance not distribute…

Our team was just a general machine learning team with no expertise in compliance issues, and only one or two of us had experience in face detection. We only cared about providing a web service that gave good face detection performance (the downstream consumer was the team that had to then take the face detection output and do business logic with it for compliance).

> amazon has the resources to scale this to thousands of RPS

Actually, this part is mostly straight forward. Not easy, but straight forward, and Amazon already has this part solved for most things. Our ML team did have expertise in making very large-scale services (our company operated a large ecommerce web store, so our traffic was actually very high, top 500 Alexa rank), but our concern over Amazon had nothing to do with whether it could handle enough requests per second.

But this also misses the point. The latency for a single request was too poor for a variety of natural images (crowd scenes, public streets, etc., -- exactly the types of images relevant for many surveillance applications). It was not a throughput issue, I guess unless you expect consumer teams to do something like chopping images up into sub-images, replacing one Rekognition request with a bunch of sub-image requests, and then post-processing to stitch the results together and account for error introduced by the choice of where to split the original image. Yikes.

It's not an issue of "scaling it up" -- it's an issue that the fundamental machine learning algorithm underpinning it suffered too high latency, likely because it was a deep neural network combined with a bunch of post-processing gadgets that are conceptually not good enough for bespoke use cases -- things which engineers on a centralized platform have little incentive to care about.

The real question is whether Amazon has incentives to make this part better, and if so, how? Will a general one-size-fits-all face detector work for customers? Or will they need application-by-application customer-specific tailored models (and if so, why aren't customers savvy enough to realize that the all-in costs to build and operate it in-house are so demonstrably cheaper than Amazon, and might be the only way to meet specific performance goals)?

I hear this so often, that Amazon can "just scale it up" -- but I think if you really work in face detection and have studied the problem deeply, you know this is not right. There's a whole different bucket of problems and issues and cost-effectiveness trade-offs going on.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#135

Earlier quoted context omitted.

> It is not inappropriate for police to be able to determine who they are interacting with if that person is a known criminal. Everyone is a criminal. You commit multiple crimes a day if you're like most people. There absolutely should be a clear delineation between what is possible and what is allowed when considering state surveillance. The 4th amendment: "The right of the people to be secure in their persons, hous…

Reasonable people can disagree on the definition of "unreasonable," however. Is it "unreasonable" for the Amber Alert system to co-opt the eyes and attention of ten thousand citizens around a kidnapped child? Hard to say. Most people seem to think 'no.' If that system could co-opt camera feeds instead of human attention, we could get most (all?) of the benefits of the existing Amber Alert without having to burn civil…

Amber Alerts are opt-out, not mandatory.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#136
post #5

At some point in time we Americans are going to have to come to grips with the computational abilities of the state. There are way too few safeguards in the system and the Bill of Rights does not seem to be adequate in dealing with the new reality that private companies can acquire a vast amount of information on us and thereby sell it to law enforcement. Law enforcement can use privately obtained information without…

> What happens with false positives? What's the recourse? The same recourse when you're false-positived by a witness, I'd assume: get a lawyer. Personally, I'd probably place my trust in facial recognition (possibly not this system specifically, though) over human eye-witness testimony, based on my understanding of human psychology and its vulnerability to incentives. It's not as dire as it's sometimes painted. MD at…

It's not the same thing as a false positive by an eye witness. Eye witness testimony is well known to be prone to false memories. Hence there needs to be corroborating evidence. I think people are too readily willing to accept a technological identification. I can see a claim being made that the computer is wrong only once in a billion cases without proper justification. There are people today who still think lie detectors are reliable.

I can see a time where the presumption of guilt is assumed simply because the technology says so. Clearly we are not there yet and possibly never will be. I hope that issues will be sorted out before the technology gains widespread use.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#137

Earlier quoted context omitted.

Let me increase your rent so we can pay for said rehab program. If you say no, then you have no right to say anything.

Are you aware of how much money it costs society to send someone to prison? A LOT more than a rehab program. - You have direct cost of imprisonment ~$60k a year - The costs of prosecution (no idea what this costs though) - Once someone has been to prison they are much more likely to be involved in crime again, and become a repeat customer of the industrial prison complex - We loose a tax payer - The perp will have no…

You haven't shown that (successful) rehab costs less.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#138

This is a ridiculous article. Rekognition has been available for over a year, and I've been using it for that whole year. I've even integrated it into my own home security camera. Amazon is selling it to law enforcement, but it's not like LE get special service that no one else does. Amazon built a generic facial recognition system, and it works pretty well. Even if they weren't pushing towards the police and governm…

I'm not sure I follow because your home security system, correct me, is only near your home, not connected to millions of people collectively that are mechanical turked unwillingly. Them apples aren't oranges.

I'm just saying that this technology is available for everyone, even in a security context. It's not like they build this just for LE.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#139
post #24

Key quotation: "Facial recognition is not new technology, but the [criticizing] organizations appear to be focusing on Amazon because of its prominence and what they see as a departure from the company’s oft-stated focus on customers." This is an interesting side-effect of the success of AWS. Amazon's customers are not just individual consumers or even sellers anymore, but organizations on the scale of multinational…

Is it me or has Amazon ever cared about customers, always marketshare and profit margin. I've never felt Amazon was a pro-consumer company, merely a side-effect that they need consumers.

I need air, so I am very air-friendly.

What reasons do you have to conclude that Amazon is not pro-consumer? Chasing marketshare and profit margin is something else, and being pro-consumer is a way to win marketshare and profit-margin.

Selling everything, generous return policies, free shipping, customer reviews; these are all customer-friendly.

Re: Amazon Pushes Facial Recognition to Police, Critics See Surveillance Risk

#140
post #31

Earlier quoted context omitted.

The number of counterfeits and abomination of UI argues against that point. I think Amazon is against both customer and small retailer and that goes to my point they are about market share and profit margin above all else.

As someone who has sold on Amazon, its not as simple as you make it sound. Amazon takes my product, which I identify, and puts it in a bin at whatever warehouse. Another seller, sends in the same product, but theirs is counterfeit. Without opening the package, how is Amazon to know which one of us is selling the counterfeit and which one is real? Now, my inventory is mixed in with amazon's and this counterfeiter, tha…

Amazon could do what Amazon used to do, and what everyone else does: Don't comingle supplies from trusted sources from random third parties. Have you every tried selling a wholesale bag of carrots to a your local grocery store? They aren't interest, for good reason.
Post reply on HN