Facebook - facial recognition is enabled | Hacker News Reader

Facebook - facial recognition is enabled

nakedsecurity.sophos.com

45 pointsrellimluap29 comments

Looking at the privacy, this was unimaginable 5-10 years ago. But Facebook's strategy of little steps is working. I'm quite sure people will accept this and most comments won't go beyond how amazing this is, creepy but still awesome and similar blabla.

Poison comes in small bottles...

This technology will theoretically enable fb to approximate your geographical location at a given time: [link redacted]

Phone carriers already have this data, and Facebook does as well every time you access their website over standard methods

> then it hit me that this picture is likely to end up on Flickr, Facebook or some other online service. And even if it’s not visible where the picture was taken, it is probably geo-tagged, and it definitely has a time stamp. Which means that there is now proof of my whereabouts yesterday, available on the internet.

One solution is to spam photos of yourself with false timestamps and geodata.

This is exactly why I don't tag photos of myself and remove tags when a friend does it. I have no idea why people are zealous about tagging people in photos, if it's a friend you already know the name, if it isn't then do you need to know?

It's not to help me identify who's in a picture, but rather to make it simpler to find pictures of a person.

Example: I want to photoshop a picture of my friend onto a dog; I could either load all pictures with him tagged in it, or laboriously go through all of my pictures from a given weekend, hoping that the picture I'm looking for is there.

Result: http://i.imgur.com/C6mrd.jpg

Personally, I blocked the ability to tag me in any photos. I don't have to remove myself because no one but me can tag me in the first place.

How did you block that?

I don't believe he/she did that, because it's not possible.

This is really amazing. With so much feedback data, they could really fine-tune this. Imagine all the other practical applications...

Imagine all the other political applications...

As I alluded to in my other post, I could only imagine the TSA getting a hold of this and integrating it into the scanners. I mean, that would be a great use for catching fleeing criminals, but of course, it's all too easy to use against normal people too as they've shown with less-advanced technologies.

If they do I will tatoo a SQL-injection attack on my forehead.

If we have nothing to hide, I think it'll work in our favor. I used to fly a lot, and at one point, I was held at US customs for 3.5 hours. In the end, apparently the reason was because I had too many different kinds of stamps in my passport.

I've also been searched numerous times (although I guess it's small, considering how much I fly). If facial recognition could help in identifying criminals (starting to sound like Lie to Me), I'd be for it.

"If we have nothing to hide,"

Right to privacy. Even people with nothing to hide don't like being monitored.

I don't see how it would help in your situation either. "Lots of stamps" is annoying to passport control regardless of any other factors (just because you're not a listed criminal doesn't mean you aren't up to something), and I don't feel like the detainment time is going away anytimes soon for these kinds of things. Even if the information is easily accessed, you have to wait for the decision-maker to spare some time to consider the info and give you the OK.

Sorry, but "nothing to hide" is not a valid reason to enact this sort of measurement. Even if a person has a clean record, someone will find a way to indict that person of a crime, and that usually happens when laws change over time. The notion of a "terrorist" and "enemy combatant" have changed over time since 9/11. A person such as yourself traveling a lot is already going to be suspect - imagine if one day you had a label like these attached to you simply because you fly a lot?

In a similar vein, one example is Megan's law, named after a young girl who was murdered by a convicted sex offender. The law was enacted to create a registry of such offenders to let people know when a sex offender would move into new neighborhoods - a modern day Scarlet Letter. The idea was "fine" in theory when it came to child molesters, but since then the definition of "sex offender" changed, which now encompasses near-age-consenting kids who would send each other risque pictures or an 18 year old having sex with a 17 year old.

There are just examples, but that's why "nothing to hide" doesn't work in our favor.

I fervently disagree. Having 'nothing to hide' most certainly means not caring whether someone is watching or monitoring you for a specific purpose. Should we let an irrational need for privacy stop major advancements in science and technology? I don't want to get too political here, but I'd like to point out I was part of a very effective stem cell cure to Parkinson's Disease (multiple years, no signs of deterioration with no upkeep), that was discontinued for legal/moral/'sense of' reasons. If you're not familiar with current treatments, youtube the surgical procedure for Deep Brain Stimulation implants. If this technology ends up being incompetent, and does not help separating good people from bad people, there is obviously no reason to implement it. However, if there is a statistically valid difference, should an absent need for privacy discredit the advancement?

Your second point is correct: the law is to be interpreted, and anyone can be convicted of a crime, if the prosecution tries hard enough. However, that's a very specialized case, and there isn't someone out to get everyone. Even being wire-tapped and watched 24/7, chances are you'll never interact with the people watching. If there's almost 0 chance of someone interfering with your life, why does it matter if your privacy is lost to them? I'm not trolling or playing devil's advocate - I really do not see why it matters. Someone I'll have 0 interaction with can watch me 24 hours a day, and it would not phase me at all. I only bring this up because you brought up terrorism. I also argued (your view) on Ex Parte Quirin several years ago, so I really have considered the subject.

Your Megan's Law example is really a tangent. I don't like child molesters, and I don't agree with statutory rape being subject to such communal ostracism - just like you. In such a case, yes, the law changing has screwed things up. However, it doesn't apply to our original point. You're arguing whether specific information should be publicly available, while I'm arguing whether already publicly available material should be analyzed. They already see your face, just humans aren't able to spit anything useful out.

I hope my point is clear.

Since so many innocent people have been charged with crimes and jailed/murdered in the past, it is wise to limit the effectiveness of systems that can come under malevolent central control with a political campaign or coup. We don't have to identify specifically who will do this to be prudent now.

You don't have nothing to hide. You just haven't realized that there's something you want hidden about your life.

I'm not a big proponent of nothing to hide on a good day, but your statement is no more than a no true Scotsman.

Very real and legitimate objections based on a right to privacy aside, the main flaw with this point of view is this: "If there's almost 0 chance..."

Almost 0 is not zero. If you're monitoring 300 million people, and the chance of a false positive is, say 0.001% per person per day, that's still 3,000 false positives every day. The false positive signal far outweighs the signal from the "real" dangerous people, making the system pointless and useless. Even if it works 100% of the time on the intended targets, human operators will become too fatigued from dealing with the false positives to deal with them effectively.

Now we can identify the child molesting TSA agents, too.

Part of me thinks it's indeed very cool, but part of me can't help but wonder when I'll walk into a store at the mall and hear "Welcome back, Artmageddon!" as some unseen Kinect-style camera looks up my face in some database shared by a whole network of advertisers.

I suppose that's about the time that this will come into fashion: http://ahprojects.com/c/itp/thesis

Google was doing work on facial recognition software, but stopped because they feared its potential negative use for surveillance in dictatorships and other oppressive countries.

Would it be reasonable to assume that facial captchas are actually more to do with building a training database than security??

No. People said the same about Google and Facebook collection phone-numbers for security purposes, but the truth is, extra authentication methods like these are crucial to recovering accounts in the event of attacks. In any case, security teams at major companies are usually at-a-distance from product teams.

In addition, companies like Facebook and Google have a distinct incentive to keep their product spam-free. This means that, so far as spam and security are concerned, the user's interests are completely aligned with the company.

I noticed this the other day. Kinda awesome and kinda scary.

Reply on news.ycombinator.com