Students? You're proposing we track students? It says something that the first example that's used (not just you, I've seen it many times elsewhere, too) is that we track the only group of people who by-and-large cannot consent, and not only that, but in a way that has permanent implications for their right to privacy.
Further, the idea of facial recognition being used in every store is atrocious. Not only can they track you online, now advertisers and the government can track your offline spending habits and location at all times, too!
Why hasn't anyone proposed a use for facial recognition that isn't:
1. Useless (like it is in iOS)
2. Completely scummy (like it is in both of your examples)
3. Using it to increase surveillance on populations that cannot legally offer consent in any way, in areas that they're forced to go to?
I kind of expect it here on HN, given how large of a proportion of adtech and other surveillance-oriented employees there are here, but I see it everywhere, and it's just confounding.
> Students? You're proposing we track students? It says something that the first example that's used (not just you, I've seen it many times elsewhere, too) is that we track the only group of people who by-and-large cannot consent, and not only that, but in a way that has permanent implications for their right to privacy.
As if we already don't track students? You just think it's somehow better if a fleshy neural net violates their consent and privacy by manually checking up on them when they are are truant because why...?
And who says everything must always be taken to the extreme? Just because facial recognition technology exists don't mean it will always abused. I'm a volunteer secretary at a local community center and I'm sick of manually performing facial recognition and updating a google doc to track attendance. It wastes an hour of my time every week. Is it so bad I want a raspberry pi attached to a camera that can auto-update the google doc for me?
There's a difference between "We're writing your name down in a spreadsheet" and the inevitable "Your facial structure is being processed by a private company that is absolutely going to share with anyone who asks" that happens with the types of companies that contract with schools.
Even assuming that they were decent companies, the Fed has shown a tendency to force companies to give up fingerprints, DNA, and so on: there's no chance that they wouldn't do the same for facial recognition data.
And who says everything must always be taken to the extreme?
Your only proposals were extremely extreme, and neither of them were particularly rare desires for this tech.
Just because facial recognition technology exists don't mean it will always abused.
Sure, like I said, there are some useless uses of it, too, like facial unlocking. What you're proposing is an abuse, however.
Is it so bad I want a [Raspberry Pi] attached to a camera that can auto-update the [Google Doc] for me?
While your proposed use, facial recognition for presumably-consenting adults, is more mundane than the initial two proposals, I spent an hour looking for an example of facial recognition that was accurate and detailed enough to tell the difference between individuals that ran on a Raspberry Pi, and came up blank. It seems as if the hardware is too underpowered to do so (or, at least, the current tooling is too bloated to), which means that you'd end up processing it in the cloud or similar, which would still mean their data was leaving your control; that is pretty bad, I'd say.
> There's a difference between "We're writing your name down in a spreadsheet" and the inevitable "Your facial structure is being processed by a private company that is absolutely going to share with anyone who asks" that happens with the types of companies that contract with schools.
I don't see why that is inevitable. My high school in 2007 had its own IT team and kept all of its data on-premises (this was before "the cloud", but still, you could just mandate that all facial recognition models have to be stored/processed on-premises).
> Even assuming that they were decent companies, the Fed has shown a tendency to force companies to give up fingerprints, DNA, and so on: there's no chance that they wouldn't do the same for facial recognition data.
Yeah, it's called warrants. I don't see what is wrong with that. If the FBI wants to know if a student was in school on a certain day, I don't see how a warrant for checking facial recognition logs is any different from checking a manually maintained spreadsheet, other than the latter being more error prone.
> I spent an hour looking for an example of facial recognition that was accurate and detailed enough to tell the difference between individuals that ran on a Raspberry Pi, and came up blank
I haven't done it yet, but this seemed good enough for my purposes[1]? As far as I can tell, that runs completely on a pi with no cloud resources...
> which would still mean their data was leaving your control; that is pretty bad, I'd say.
Even if I needed to use cloud compute, I take issue with your phrasing. I don't believe people "own" information about themselves. If we are both in public, and I take a picture of you, write down facts about you, or otherwise observe you, I have not stolen anything of yours, and I do not need your consent to have done so.
The annoying thing about the emphasis is that there is an obvious use for it that won't suffer from an overload problem - personal usage for those of us who have terrible face memory/recognition and ability to manually enter/assign a string to a face.
That would combine with AR. But that AR has already flopped due to several issues even without the added camera needed. Plus the issues of existing eyeglass combining.
Really that Glass backlash was bizzare along with the post 9-11 "anyone pointing a camera near anything man made must be a terrorist" yet nobody bats an eye or does more than snark at surveillance cameras everywhere. There is probably some psychological principle behind it but the simplest explanation is that "normal people" are fundamentally fucking insane and only care about it conforming to norms instead of actual harm.
The Glass backlash was for the wrong reasons, but right overall. There are plenty of individuals who've made their own setups for this, and setups that don't share with Google, at that. Glass, though, was definitely harmful.
I hate to cross-reference as that's technically breaking the rules, but their post history doesn't exactly suggest that they're being sarcastic, and their (three-person, all related) company's privacy policy seems to be "We're taking your data, and hell yes we're going to sell it."
The privacy policy I have on my website is used for exactly one app in the Google/iOS App store and is based on a template one I found on GitHub. FWIW, that one app collects exactly zero data on users other than whatever Unity collects for serving opt-in ads.
Further, the idea of facial recognition being used in every store is atrocious. Not only can they track you online, now advertisers and the government can track your offline spending habits and location at all times, too!
Why hasn't anyone proposed a use for facial recognition that isn't:
1. Useless (like it is in iOS)
2. Completely scummy (like it is in both of your examples)
3. Using it to increase surveillance on populations that cannot legally offer consent in any way, in areas that they're forced to go to?
I kind of expect it here on HN, given how large of a proportion of adtech and other surveillance-oriented employees there are here, but I see it everywhere, and it's just confounding.