Borders Nixes Face Recognition

jeffy124 writes "Due to recent criticisms surrounding their implementation of face-recognition technology to watch known shoplifters, Borders Bookstores is suspending the approach. This doesn't mean it's gone for good, it may return in the future. They want to resolve the issues brought up by privacy and human-rights activists."

Comments

Trial balloon management

[sllort]

Another example of "Trial balloon management".

The formula:
- We'll announce that we're doing something, but only introduce it on a low cost basis into a small target market.
- We'll watch the reaction.
- If it's bad, we'll denounce ourselves and retract our low cost trial balloon.
- If it works, we'll exploit the hell out of it.

This formula has been applied with both results to:
- SmartTags
- Windows Activation
- Borders Face Recognition
- Skylarov
- Implementation as a "Trade Secret" (ms & kerberos)
... and on & on.

Other examples?

Re:I don't understand...

[dragons_flight]

Why do large companies like Borders announce implementations of things like this, suspend them upon complaints and then review things like customer's rights to privacy? Are these only an issue when people complain?

Why not? It makes good business sense. Lots of places have security cameras, no one really would have cared if that's all they wanted. I have no idea how much they lose to shoplifting but it might be enough to financially justify installing such a system. From their point of view they are just protecting their possessions from theft.

Clearly someone knew that people might be upset by this, otherwise there is no point in announcing it, you just start doing it. Instead they sat down, told people what they wanted to do and waited to see the reaction. Now they've realized that it isn't a reasonable thing to do unless they can seriously reassure the people of their privacy.

I bet we still see systems like this appear, but it isn't a place like Border's that will likely stand up and take the intial flak. Perhaps casinos, banks, or some other place where security truly matters will be the first.

privacy?

[amoken]

What does this have to do with privacy? Borders is a corporation, and its property is private property. If they want to implement something like this, it's fine to complain to them on various grounds, such as that the technology can't be trusted (as though a person could), but to attack them on privacy grounds is absurd. If someone said you couldn't enter their house without being photographed or under video surveillance or whatever, would you attack them on privacy grounds, or would you just leave and tell them they were being silly? It is not your right to shop at Borders.

Re:privacy?

[sludgely]

Borders is entitled to do whatever it wants inside its own stores, but the consumer does not have to stand for it. Also, if what they are doing leads to descrimination or if there are mix ups like there have been in the past, it could most likely lead to problems. Most people are afraid of what this technology can lead to and will therefor shop elsewhere. Borders needs to decide its priorities.

Re:privacy?

[NMerriam]

What does privacy have to do with government? Is it not possible to have privacy from a corporation?

Re:privacy?

[ClipDude]

What does this have to do with privacy? Borders is a corporation, and its property is private property.

Just because a corporation is involved, and it involves that corporation's private property, doesn't mean there aren't privacy issues involved. Let's pretend my school installed a secret camera in my dorm's bathroom and videotaped me showering. Of let's say they put in my room and recorded my conversation with my girlfriend. Either of these, would they to occur, would involve a (not-for-profit) corporation doing something on its private property. It's not my right to go to school there (as the admissions department reminds unlucky applicants every year). But if my school did either of these things, I would be quite upset, and would consider them to be an invasion of my privacy.

Let's think of it another way: Do you shop at Safeway (or any of their subsidiary stores)? Do you use their Club Card? Now, what might have you purchased at Safeway that you wouldn't want the world to know? Condoms? Birth control pills from their pharmacy? Anti-depressant medication from their pharmacy? Hemorrhoid treatments? If you have purchased any of these things, and use a Club Card, it's probably in their database. A private corporation chronicling what occurs on their property. But you would be upset if they put that on their web site, wouldn't you?

Re:Privacy and the lack there of in the World toda

[Mr. Frilly]

eh, good point on the home burgler thing, except maybe I'd have my own at home video system :)

Your credit report analog, though, is by far the best reason I've seen yet on this forum as to why we should be concerned about these systems. Then again, I think credit rating report systems are a good thing, and believe me, I've seen plenty of my friends get screwed over by these things. But in the end, the problem really isn't that the credit report exists, it's that there's no good system for removing an incorrect entry.

I think what these video recognition systems really need, is a legal incentive to insure that the cost of a false positive is very high. That way, it would be the burden of the seller to ensure that their databases/reports are correct, unlike the way it currently is with credit reports.

Re:Police database

[Chasuk]

I really don't see the big deal.

Let's look at it from several perspectives.

First scenario:

Suppose that I am an employee at a large department store. I've worked there for many years, and I've been present during the apprehension of shoplifters who were arrested and convicted of their crime. One day at work I see on of these former perpetrators enter the store. Would it be unreasonable for me to monitor their activities with especial care?

Second scenario:

There is a concealed room above the same department store where a team of observers sit, watching the shoppers as they enter. These observers have photos of known shoplifters taped to the walls of their hideaway, and they are comparing the faces of the all of the patrons (honest and dishonest) with the faces of the known perpetrators. Is this unreasonable?

Addendum to scenario 1:

I follow a customer one day because I believe that he/she is a former shoplifer, only to realize that I was wrong, and that my facial recognition skills are not as acute as I had imagined. The store detective (who I notified) wasted half an hour.

Addendum to scenario 2:

One of the trained observers in the hideaway erroneously identifies an honest customer as a shoplifter. The store detective wasted half an hour.

Scenario 3:

Automated software performs the same comparisons that the trained observers performed, only it does this with extreme efficiency. It occasionally makes mistakes, and on this occasion it made a mistake on the same day that I did. The store detective wasted another half an hour.

Final proviso:

In each of the these instances, no arrests were made, because none of the suspects were observed taking unpaid merchandise from the store.

I ask: does the method of observation matter?