141–150 of 201 posts
Out of topic question. Would it be possible to combine machine learning with something like 23andMe and match 100 000 person photos with their DNA. Would we have an AI model which basically can tell how the persons looks like from a strand of hair? Imagine finding some DNA on a crime scene. Witnesses can say that this person was a bit overweight so you slide some slider to the right a bit and get a pretty good AI gen…
And oh so biased
This thread is caught up in semantics, which is regrettable The thesis of the article should have been articulated "AI is disproportionately used as a surveillance technology owning to its computational requirements, the current state of industrial-user relations and the inability for legal systems to respond fast enough to protect individuals" Edit 1: I don't know how to make that more consise without losing the mea…
Is it really inability when not doing so benefits you and your lobbying friends?
I can run models locally on my laptop, so no, it isn't. I can even train it locally on my own data if I want to. Hyperbole really serves no purpose when it's so extreme. Instead of talking about the merits of "how can this technology be abused?" you're putting everyone else on the defensive, having to say "I'm not abusing it". Like, > AI is a way, I think, to entrench and expand the surveillance business model,” she…
Which is irrelevant if tens of thousands of other AIs, from entities with 4-5 orders of magnitude more times the resources and power over the economy/society/country than you use them for surveillance and to optimize to your detriment.
The claim was never was about the logical or practical impossibility of AI to also be used otherwise.
And facial recognition is just the tip of the iceberg of AI uses in surveillance.
Nor is "blurring out faces" much consolation. What's blurred in, say, a public "Street View" service doesn't mean is also blurred on the same company's raw captured data. And the example wasn't about such cases, but about surveillance cameras anyway - which don't blur anything, just use the AI to identify targets.
We should instead of arguing the point (and especially, the point at large), instead of nitpicking on the speaker's expression, while onstage, during a conference panel talk - as if those would be the best formulated and exhaustive.
Personally, I place some hope in open source, open LLMs, self hosting AI, etc. I say “some hope” because the surveillance industrial military complex has so many more resources than organizations like Hugging Face.
I can run models locally on my laptop, so no, it isn't. I can even train it locally on my own data if I want to. Hyperbole really serves no purpose when it's so extreme. Instead of talking about the merits of "how can this technology be abused?" you're putting everyone else on the defensive, having to say "I'm not abusing it". Like, > AI is a way, I think, to entrench and expand the surveillance business model,” she…
The term "surveillance business model" itself is a serious abuse of language. Surveillance is something done by states and police forces, usually to catch criminals or suppress dissent. People use it as a label for Google's business model simply because it has a negative connotation. It's a cheap tactic in my opinion.
This has no more validity than "the internet is fundamentally a surveillance technology." It's a conflation of a tool's sometimes harmful usage with an inseparable aspect of its nature. And the idea that AI "requires the surveillance business model" is demonstrably false. Thousands of researchers over the last 7 decades have contributed to the development of the collection of technologies we call AI without that busi…
I mean at some point the conflation itself is rather fair game, given that such massive swaths of the modern Internet apply it directly. The most predominant search engine sells ads. The most predominant social media sites, all of them, run on paid placement of content (ads). In fact, every social media site with even a little traffic besides perhaps Mastodon does (I honestly don't know I've never used it). Twitter/X is trying to get off advertising as it's revenue source and because of that is worth a 10th of what it was when Elon bought it less than a year ago. Amazon you could make an argument isn't based on ads (entirely) but it's also the largest retail website in the history of man and that's probably the only reason, and also, it's largest cash cow at the moment is AWS which is absolutely shoulders deep in ads, ad services, serving ads, and ad targeting.
Podcasts? Ads. Listicles? Ads. Substacks? Ads. Need a driver file to make an ancient videogame work? You'll find a website that hosts it, alongside some ads. Wanna download some ROMs for SNES games you definitely own? Ads. Pirating some pornography with surprisingly good production value that's riffing on Pirates of the Caribbean? Sounds fun, also, the website you get the torrent link from has ads.
And like, you can get into the semantics of there's no other way to fund a website, they didn't all choose it, some don't use surveillance to serve their ads, and all the rest, and to that I say, you have much more trust in corporations than I do. Salesforce would sell my grandmother into sex slavery if they could make $5 doing so so you'll have to forgive me if I just don't necessarily trust the internet ad machine when it says pinky promise we aren't exploiting your data, especially since every time they have a data leak incident, oops, there's all my data again. Aw shucks aw geez.
User data is practically the only currency on the internet anymore, and user attention in turn is the only resource being mined by social media. Of course it's for advertising. What else would it be for? What else could it be for?
> However this is Meredith Whittaker so we can assume she knows that. So yeah, those are provocative statements made for rhetorical effect to emphasize that power will use technology to do bad things and enshrine and enhance its power.
I mean that seems like something worth emphasizing? Ad tech has had a massive influence over the modern internet space, which is why it's so fucking bad. Basically every terrible decision you lament on a daily basis, every dark pattern, every annoying thing, can in some way or another be traced to ad tech. All that shit about when you talk about needing a new hose near your phone and suddenly you have ads for hoses, that's ad tech. The entire pivot to video fiasco! Facebook wanted to stream video and bullshitted publishers and media orgs with cooked numbers that it was "what people wanted" even though demonstrably no they fucking did not, do not, and never will want that, it's what ad tech wanted and therefore it was sacrosanct, right up until it killed literally hundreds if not thousands of websites. Except even that wasn't enough, because ad tech still wants it, and so we have video platforms left and right because these companies do not work for us, they work for advertisers.
Like no it isn't everything, no single entity could be that bad. I'm just saying if all of ad tech was Thanos-snapped out of existence tomorrow, the Internet would be incredibly different, better in a lot of ways, so I feel the conflation is fair game.
This thread is caught up in semantics, which is regrettable The thesis of the article should have been articulated "AI is disproportionately used as a surveillance technology owning to its computational requirements, the current state of industrial-user relations and the inability for legal systems to respond fast enough to protect individuals" Edit 1: I don't know how to make that more consise without losing the mea…
I can run models locally on my laptop, so no, it isn't. I can even train it locally on my own data if I want to. Hyperbole really serves no purpose when it's so extreme. Instead of talking about the merits of "how can this technology be abused?" you're putting everyone else on the defensive, having to say "I'm not abusing it". Like, > AI is a way, I think, to entrench and expand the surveillance business model,” she…
The term "surveillance business model" itself is a serious abuse of language. Surveillance is something done by states and police forces, usually to catch criminals or suppress dissent. People use it as a label for Google's business model simply because it has a negative connotation. It's a cheap tactic in my opinion.
This reads like parody.
Let's say that states and police forces alone (for the sake of the argument) used to do surveillance.
For the last 2 decades, Google, Facebook, and co ALSO collect data in 1000x the times the data breadth and detail than even Stazi was capable to collect, including collecting data covertly, and abuses it for all kinds of purposes, both aggregated and targeted.
Not to mention that, beyond the surveillance they do for their private interests and business, even the "states and goverments" vs "Google and Facebook" are not even remotely separate from surveillance by states and police forces, cooperating, sharing data, or even ordered to give surveillance data, to both.
This thread is caught up in semantics, which is regrettable The thesis of the article should have been articulated "AI is disproportionately used as a surveillance technology owning to its computational requirements, the current state of industrial-user relations and the inability for legal systems to respond fast enough to protect individuals" Edit 1: I don't know how to make that more consise without losing the mea…
Technology develops as the result of focused application of capital and shaping the evolution of that technology to serve the interest of that capital. The driving force behind AI -- when ranked according to impact and significance -- is being funded by surveillance capitalists. In that sense one can say "AI is fundamentally a surveillance technology".
AI and Capitalism are a toxic mix, in my opinion.