4 ms·
<puts on tinfoil hat> Does anyone think privacy of any real sort is maintainable going forward? Machine learning algorithms are learning to identify people jus
by gcanyon 3y ago
<puts on tinfoil hat>
Does anyone think privacy of any real sort is maintainable going forward? Machine learning algorithms are learning to identify people just by their walk -- no face recognition required. Algorithms are moving toward being able to decipher text just by the audio of the keyboard being typed on.
In short, given a gestalt of ALL public data and sufficiently advanced algorithms is there really a way for people to maintain what we today consider reasonable privacy without extraordinary measures, unfailingly applied?
To be clear, I'm not value-judging the situation, just expressing what I think the ongoing trend is.
- wimp 3y agoPrivacy is a policy issue, not a technical issue. We need to focus on advocating for more useful and effective privacy protections as citizens, instead of focusing on technical evasion strategies. Because as you're pointing out, that is a losing strategy in the long run.
- dylan604 3y ago> Does anyone think privacy of any real sort is maintainable going forward Probably not, but it doesn’t mean we can’t guide the conversations about how it looks in the future. Sitting idly by just means they win, but discussing it in the open means that we might be able to put some safe guards in place. Oh, who am I kidding. We’re all screwed and evilCorp will win so we’re just wasting our energy and making ourselves crazy fighting. Resistance is futile
- pembrook 3y agoAgreed, but also "privacy" is an abstraction that layers over the actual thing that people are worried about. Any answer to why do you care about hiding this information? can all be boiled down to the fear that "[person or group] might use [private data item] to create [bad outcome] for me." So the thing people actually care about is the risk of bad outcome, not the actual data itself. If your theory is correct, then the focus should be on the prevention of asymmetric power imbalances in societal transactions that can even create [bad outcome].
- acuozzo 3y ago> the focus should be on the prevention of asymmetric power imbalances in societal transactions The rules governing social systems built to obscure the jungle (e.g., political, legal, and penal systems) can always be trumped by that which they were chosen to tame. This is the unfortunate reality of our wetware. > the thing people actually care about is the risk of bad outcome, not the actual data itself "Bad" is subjective, no? Is it good or bad if a father learns that his teenage son is not his own?
- nulcow 3y ago> "Bad" is subjective, no? > Is it good or bad if a father learns that his teenage son is not his own? I feel like that would depend on the person. If the father wanted to know and the son didn't, that would be good for one and bad for the other, and vice versa.
- berniedurfee 3y agoI don’t think that’s necessarily true. I don’t think privacy is an abstraction, but rather a very real and visceral concern. Regardless of the outcome, I’m entitled to my privacy and my privacy is important to me simply for the sake of my autonomy. Simply having my privacy violated is a bad outcome in and of itself. Relinquishing privacy in order to participate in modern society in a meaningful way is one of the great frustrations of modern times… for some.
- titzer 3y ago> So the thing people actually care about is the risk of bad outcome, not the actual data itself. It's more than that. Privacy affects the psychological context for daily activity and alters behavior, sometimes subtly, sometimes overty. If a person knows they have no privacy, they will go about life completely differently, and they will think about life completely differently. Privacy is a freedom-of-thought, freedom-of-action, and also freedom-from-anxiety consideration.
- zamadatix 3y agoEven taking privacy as a pure abstraction over things that people really care about at face value this conclusion bakes in a ton of assumptions. The assumption a perfect power balance can be attained, the assumption once attained it can be guaranteed to never falter, and the assumption what happens between now and when such a utopia is achieved is an irrelevant concern to the individual at risk now.
- agnosticmantis 3y ago> Machine learning algorithms are learning to identify people just by their walk... Turns out the UK government was working on privacy-preserving walks decades ago: https://youtu.be/eCLp7zodUiI https://youtu.be/eCLp7zodUiI
- AlwaysRock 3y ago[dead]
- duped 3y ago> Machine learning algorithms are learning to identify people just by their walk -- no face recognition required. About a decade ago I knew people researching computer vision algorithms doing non-facial recognition (stuff like ear shape/gait/etc) because companies like Fortinet were trying to build "automated doormans" to apartment/condo complexes where they would scan and analyze any humans walking by the cameras placed at the door. Not a lick of ethics from anyone involved. What we need is rabid legislation that encodes a right to be forgotten, because clearly an expectation of privacy isn't enough. I don't think there's anything inherently wrong with automating identification, but I do think there's a lot wrong with companies trying to do it for every human being that they can possibly find without any consent.
- zlg_codes 3y agoYes, it's absolutely possible. What's not possible is the society respecting privacy. As a whole, nobody respects it anymore, some even engage in half-assed devil's advocating about it. People are easy to mislead and so that's what's been done. In the future, privacy will have to be enforced through jammers and Faraday cages. None of the gait and keyboard detection attempts work in field conditions.
- StopHammoTime 3y agoI personally don’t. I used to lose my mind over the thought of my confidential documents being leaked. Then after seeing how poorly personal information is handled, I realised it’s almost a guarantee. A few things from Australia (which has good privacy laws) that made me recognise the futility of it all: 1) the large hack of Optus in which about half of the population had their credit card details stolen. 2) the large hack of Medibank in which the details of a large portion of private health insurance customer details were stolen. 3) I applied for a mortgage and found out every 2-bit mortgage broker is emailed 100s if not 1000s of sensitive ID documents every year and they definitely do not go through their email and delete them after the closure of deals. 4) Most companies in Australia only require a name, address, and, birth date to verify identity which is easily found with five minutes of searching most of the time. 5) I set up a pin with Telstra that should have blocked administrative changes on my account for years. One day I called in, got my password ready, and they didn’t ask for it. They just did it anyway. It was entirely futile. IMO the only way that privacy will ever become respected is if we move the onus for fraud onto the actual victims of fraud: the companies. This is the whole ancient joke about someone’s identity being “stolen”. It wasn’t stolen, your verification procedures ultimately failed as a business and you are trying to divert responsibility to avoid having to suffer a loss. This is one of the reasons I use my credit card exclusively these days - if it used fraudulently I know that I can charge back, and that’s about the only mechanism I can use to truly prevent unauthorised access to my money.