Until our way of governing ourselves doesn't change, everything scientists and engineers do can and will be converted into a weapon or tool of oppression. We should be more vigilant and we should demand better from our elected officials.
Well, the one edge of this sword is essentially being used by wealthy, businesses and those in power to gain additional information on average citizens to leverage for their own advantages and the expense of those folks happiness.
Traditionally, an average person had a tremendous amount of privacy simply by the virtue of being average. It wasn't worth the time/effort/resources to collect data on them. With ubiquitous sensors, it's now far more manageable or even trivial to collect data on people and use it no matter how unimportant or insignificant you are in respect to money/power.
People of great wealth and power also somewhat enjoy this privacy. Unless you're a public facing official or on the top of the wealthy list, you essentially can also disappear in crowds and enjoy a degree of anonymity. Some may recognize Gates or Bezos in a crowd but drop down to say Forbes 400 rank... #144 (picked at random) and you'll find.. (searching) Leonard Sterm.
Prior to that sentence I've never heard that name and certainly wouldn't recognize him in public even though he is estimated to be worth a modest $4.8B. I might (probably not) interact with them in the future and as an average person, I'm at a disadvantage because I can't recognize his net worth and public information which I could potentially leverage to my advantage in the same fashion someone wealthy/powerful may use their resources to leverage me.
Take this a step further and focus on business investors and participants promoting these technologies, pouring money into them. For example, I can't even name the people involved in Cambridge Analytica even though they were in the news and probably wouldn't recognize them either. Sure, I can dig around and find them but that wouldn't be useful in most situations.
For the most part, your average person doesn't have knowledge or resources to access and leverage such data that businesses are starting to use against the average consumer and worker rapidly and in ways that significantly impact their quality of life.
As such, I would propose leveraging these same identification and recognition approaches against those pulling the strings or in public government office positions who should be fighting for citizens rights and developing a targeted dataset for those folks and exposing it in a way your average citizen could use it. Perhaps as a phone app that can pass a few frames of sampled video back to a service that identifies these individuals and gives detailed public information about them. Make their anonymity suffer. Link it to public facing records of their history, etc. Develop network information for their families, friends, business associates.
Obviously, this is worthless against high profile individuals (celebrities, high government officials, etc.), but a lot of wealth and power hides in the shadows of the general public's eyes yet are pulling the strings on these cultural shifts us plebes are dealing with. Turn the tables and make the technology problematic for their daily lives as well--drag them out into the light for the public--maybe then you'll see more regulation and penalties for using these technologies surface. Just a thought.
I think you've got something there. This would likely need to leverage decentralized tech (perhaps blockchain could actually be useful for something other than crypto-currency?).
Then there are the challenges of getting people to pay attention and to care about the results.
Just tested it, it returns people with some similar features but that are obvious different persons, different face shape and size , maybe facial hair and glasses confuses it or they could not find any better matches so they returned just something so it is not empty.
I tried it with a picture of me and the top half dozen photos were, in fact, pictures of me. I was originally thinking there were also bizarrely a couple of photos of signs and then I realized I was standing next to a sign in the photo and that's what caused the algorithm to stick a couple of signs in there.
Amazing! I just tried it on a picture of me, and it found several other photos of me that I had no idea were on the Internet, and a few photos of people that looked so much like me that I thought it was until I took a close look.
From what someone said in a previous thread the copyright part is not that simple, so as an example the person that takes the picture can claim the copyright not the subject. Probably there is a different law that would prevent others to make money using your image but is not the copyright AFAIK
But selfies would be pretty unambiguous, no? Maybe even enough to get the ball rolling on a class-action, if enough Californians requested their data and got back a bunch of selfies.
I'd also like to see what would happen if Europeans reported them to their data protection agencies. Nothing will happen to the company, since it's in the US, but it might make the life of the execs hard if they ever wanted to do business in Europe.
>Probably there is a different law that would prevent others to make money using your image but is not the copyright AFAIK
There is a right of publicity [1] that tends to come into play if I want to use a photo of you for commercial purposes
like advertising. However, if I sell a candid photo of you that I took in public to a newspaper or if it's in a book I sell that's OK. (Assume it's straightforward use and nothing that could be misconstrued or deliberately misrepresents you in some way.)
No idea about this case and IANAL. But it seems in general that I would be able to freely use images on the Internet for training models and for other purposes that don't involve republishing those images.
I'm trying to understand where subject release documents fit into this. So I don't have to get those signed to take pictures of people, even in public? What if it's a still from video?
A model release form gives the right to use a photo for most purposes* including things like marketing literature, e.g. the classic happy smiling diverse workers found on company websites and in promotional brochures. However, if I just take a photo of you in a public place I can upload it to Flickr, post it on my website/blog, sell it to The New York Times for editorial use, etc. You have fairly limited rights to photos of yourself taken in public.
A video is just a sequence of still frames so I don't know why that would be any different.
IANAL. Added: I'm primarily familiar with the US. Your mileage may vary elsewhere.
*subject possibly to what I wrote earlier about misleading or defamatory purposes, e.g. if I use a photo of you to illustrate an article about neo-nazis.
But image similarity is not the same as face recognition. I just uploaded the photo the author gave to Clearview, and it returned none of the photos Clearview found – just other front-facing flatly-lit mugshots of other people.
Of course Google can do this internally (and FAcebook could do it even better) – but the public-facing image search does not do what Clearview purports to do.
Exactly! People think the Internet is like Disneyland but it's actually more like the bad parts of Bangkok.
Somewhere between the inter-networked machines and peoples' minds some kind of magical thinking takes over.
Why don't normal people understand that they are being actively fought by organized relentless enemies?
In the old days cities had walls around them, I think we're barely getting to that point on the internet now, despite repeated and continuous raids by all kinds of marauders.
Or the digital world is an incredibly huge and complicated place that most people don't even begin to grasp. What is private and public is not obvious, and everything can quickly leak, be accidentally shared, ... And you lose control.
Stop blaming the users for what is systemic and legal faults - and in many cases really an issue not even of laws or of enforcement. Think revenge porn - nothing you can do to stop it, and only so much you can do to contain it once it's out there.
Look I'm not saying it's fair, I'm saying that, if you want e.g. control of your own nudie pics, you got to (excuse me) come correct. Don't take them on your phone, and don't ever let them onto a computer that's connected to the internet. Better yet, don't use digital at all, use film and keep it in a safe, like everyone did before pocket computers with radio networking and hi-res camera became ubiquitous.
There are things you can control and things you can't, eh?
I do blame the users. They buy a book of matches and proceed to burn their own house down. Either they're children or just incredibly stupid. That everybody else is doing it too changes nothing IMO.
Maybe just add noise. For example, there are lots of apps that mix faces. So mix your face with a bunch of other faces, or even faces from https://thispersondoesnotexist.com and upload the results wherever.
What astounds me is that GDPR seems to have an actual effect. When it first became law I just assumed that companies would just move their servers to $not_europe and create shell companies if required.
> "Clearview does not maintain any sort of information other than photos,” the company wrote. “To find your information, we cannot search by name or any method other than image."
So a police force buys this product, searches for an image, and just gets more images. Really? What use would that be?
I like the way they claim it stops "child molesters" but then goes on to say that Macy's uses it, presumably to identify shoplifters before they strike.
>> “Clearview helps to identify child molesters, murderers, suspected terrorists, and other dangerous people quickly, accurately, and reliably to keep our families and communities safe.”
It helps to identify child molesters. "Think of the children".
It's worth noting (as mentioned they are under security review) that apparently ClearviewAI lost their client database (hopefully that's all) to hackers last week. Along with being sued by the big boys, that's gotta hurt.
Is it just me or are those results clearly based on using her name (from the ID she provided as part of her CCPA request)? To have zero false positives, and multiple different hair colors, etc ... strains credulity that is all based on face matching only.
Supposedly any company in the EU that has data on you needs to proactively make you aware of this and allow you to have this data erased or rectified.
I don’t think all or even most companies respect this rule yet (I haven’t heard of a single company about my data so far) but I think it’s a powerful idea that would keep companies like Clearview from being able to build up such large databases without people knowing about it.
I really hope the EU will keep strengthening subject rights as I think it’s a very good solution to most problems involving the use of personal data.
44 comments
[ 2.9 ms ] story [ 99.8 ms ] threadIt would be nice to find ways to take these weapons and use them for self-defense as well. I wish I had a clue of what that would look like.
But good luck with that; there's a reason why that sort of information isn't sold on a per-person basis.
Traditionally, an average person had a tremendous amount of privacy simply by the virtue of being average. It wasn't worth the time/effort/resources to collect data on them. With ubiquitous sensors, it's now far more manageable or even trivial to collect data on people and use it no matter how unimportant or insignificant you are in respect to money/power.
People of great wealth and power also somewhat enjoy this privacy. Unless you're a public facing official or on the top of the wealthy list, you essentially can also disappear in crowds and enjoy a degree of anonymity. Some may recognize Gates or Bezos in a crowd but drop down to say Forbes 400 rank... #144 (picked at random) and you'll find.. (searching) Leonard Sterm.
Prior to that sentence I've never heard that name and certainly wouldn't recognize him in public even though he is estimated to be worth a modest $4.8B. I might (probably not) interact with them in the future and as an average person, I'm at a disadvantage because I can't recognize his net worth and public information which I could potentially leverage to my advantage in the same fashion someone wealthy/powerful may use their resources to leverage me.
Take this a step further and focus on business investors and participants promoting these technologies, pouring money into them. For example, I can't even name the people involved in Cambridge Analytica even though they were in the news and probably wouldn't recognize them either. Sure, I can dig around and find them but that wouldn't be useful in most situations.
For the most part, your average person doesn't have knowledge or resources to access and leverage such data that businesses are starting to use against the average consumer and worker rapidly and in ways that significantly impact their quality of life.
As such, I would propose leveraging these same identification and recognition approaches against those pulling the strings or in public government office positions who should be fighting for citizens rights and developing a targeted dataset for those folks and exposing it in a way your average citizen could use it. Perhaps as a phone app that can pass a few frames of sampled video back to a service that identifies these individuals and gives detailed public information about them. Make their anonymity suffer. Link it to public facing records of their history, etc. Develop network information for their families, friends, business associates.
Obviously, this is worthless against high profile individuals (celebrities, high government officials, etc.), but a lot of wealth and power hides in the shadows of the general public's eyes yet are pulling the strings on these cultural shifts us plebes are dealing with. Turn the tables and make the technology problematic for their daily lives as well--drag them out into the light for the public--maybe then you'll see more regulation and penalties for using these technologies surface. Just a thought.
Then there are the challenges of getting people to pay attention and to care about the results.
Other search images can obviously perform this function, but I think most stay away from it because of the creepy factor.
[0] https://yandex.com/images/ - click on the camera for the reverse image search
I'd also like to see what would happen if Europeans reported them to their data protection agencies. Nothing will happen to the company, since it's in the US, but it might make the life of the execs hard if they ever wanted to do business in Europe.
There is a right of publicity [1] that tends to come into play if I want to use a photo of you for commercial purposes like advertising. However, if I sell a candid photo of you that I took in public to a newspaper or if it's in a book I sell that's OK. (Assume it's straightforward use and nothing that could be misconstrued or deliberately misrepresents you in some way.)
No idea about this case and IANAL. But it seems in general that I would be able to freely use images on the Internet for training models and for other purposes that don't involve republishing those images.
[1] https://www.nolo.com/legal-encyclopedia/the-right-publicity....
A video is just a sequence of still frames so I don't know why that would be any different.
IANAL. Added: I'm primarily familiar with the US. Your mileage may vary elsewhere.
*subject possibly to what I wrote earlier about misleading or defamatory purposes, e.g. if I use a photo of you to illustrate an article about neo-nazis.
Just use their reverse image search feature, as helpfully explained here:
https://support.google.com/websearch/answer/1325808?co=GENIE...
They have tons of images from social media sites, too.
Of course Google can do this internally (and FAcebook could do it even better) – but the public-facing image search does not do what Clearview purports to do.
California/GDPR laws are great but they don't stop the data collection, merely allow the user to access it.
We need to make companies like Clearview completely and unquestionably illegal. I am infuriated and don't know what we can do.
Somewhere between the inter-networked machines and peoples' minds some kind of magical thinking takes over.
Why don't normal people understand that they are being actively fought by organized relentless enemies?
In the old days cities had walls around them, I think we're barely getting to that point on the internet now, despite repeated and continuous raids by all kinds of marauders.
Stop blaming the users for what is systemic and legal faults - and in many cases really an issue not even of laws or of enforcement. Think revenge porn - nothing you can do to stop it, and only so much you can do to contain it once it's out there.
There are things you can control and things you can't, eh?
I do blame the users. They buy a book of matches and proceed to burn their own house down. Either they're children or just incredibly stupid. That everybody else is doing it too changes nothing IMO.
So a police force buys this product, searches for an image, and just gets more images. Really? What use would that be?
The response report on the author, for instance, clearly identifies her
It helps to identify child molesters. "Think of the children".
https://nakedsecurity.sophos.com/2020/02/28/clearview-ai-los...
I don’t think all or even most companies respect this rule yet (I haven’t heard of a single company about my data so far) but I think it’s a powerful idea that would keep companies like Clearview from being able to build up such large databases without people knowing about it.
I really hope the EU will keep strengthening subject rights as I think it’s a very good solution to most problems involving the use of personal data.