Meta has announced this week that it's launching A.I. products across Facebook (and I assume Instagram, which I don't use). Several people from different quarters have asked me about it, so rather than repeat myself I've blogged about it.
https://blueteamhackers.com/illegitimate-consent/
Hope it helps.
Interesting read.
I'm very concerned about how AI is creeping into so many parts of our lives with little regard to anything other than the huge profits to be made.
To me, it's one facet of the perfect storm humanity seems to be intent on creating for itself this century.
Hope it helps.
Ta ,
Never had an account whether Facebook/instagram but I've forwarded it onto a mate and I'll be dealing with mums Facebook t'morn.
Facebook is pure filth.. I was getting pictures from a user called 'GETFUKD' on my timeline, I won't post the images, but I reported them for 'hate speech'.
The reply from facebook was, and I quote, "we didn't remove the photo" because it didn't violate the standards or something.
Thanks! I am in two minds. I don’t want AI trained on my images (I think they feature my friends and relatives and I’m not second guessing whether they want to be ai fodder) but I am less sure about words. I like to think that ai like me would be more reasonable and informed and less prone to hallucinations…
Just for context I occasionally press LinkedIn’s ‘rewrite with AI’ button and get my words turned into anodyne marketing speak. I haven’t used it. I would rather be slightly rubbish and real.
already opted out this morning at 5 am!
When do we find out if we've won and managed to get removed from it?
Appears to be within 2 minutes...my objection has been noted and will be carried forward. Unsure it'll stop them doing anything, but it was nice of them to play the game!
Just opted out. Just said I do not give permission. Instant reply saying they'd honour my objection. I didn't give a reason for it.
Is there a way to opt out in the app?
Yep, instant and presumably automatic agreement. No idea if I'd know whether they did or not.
@cougar: just read your blog. Very good. Like the humour!
How about another post about how to deal with our supposed life partners (and joint account holders) who fail to take security seriously? 😑
To object via the App
Help Centre
Search "object AI"
"Information about objections ... "
Then follow your nose as the path gets easier. The actual link to the objection form is a hyperlink saying "here" in the middle of a paragraph of text.
Hope this helps others.
I don't think child accounts are affected but if anyone knows different maybe post in here. 🙂
Thanks Cougar.
If anything is going to send AI over the edge, it's going to be facebook 😕
It’s not something I’m going to get too concerned about. I fully subscribe to the principle that if you put any information or data onto a social media platform you should expect them to do stuff with it. If you don’t want it used, don’t give it to them.
You are the product.
The link from the blog worked for me fine- but not when I forwarded it to Mrs K. Strange
Thanks Cougar
Done , thanks 👍
Not just AI - I hate the way simply turning off cookies has become the same as reading 48 pages of T&Cs!! It’s a blatant way of confusing people into just clicking accept for an easy life.
A big - “just sod off” button should be made mandatory.
I’ll read your piece later Cougar.
talking of which we seem to have acquired a floating privacy tab!!!
<<<
If anything is going to send AI over the edge, it’s going to be facebook
Dunno. Microsoft Total Recall is going to do a pretty good job too.
You are the product.
In the case of Facebook, yes. Even if you don't even have an account. Back when they introduced all that tagging nonsense, they believe what others tell them about you more than what you tell them.
In the case of Microsoft. Even if you pay and buy a licence, you are the product. And now they have spyware "but it is encrypted".... err yeah with keys that they look after for you.
NSA and Chinese govt must be rubbing their hands right now. Who needs a backdoor when they can put in a front door and keep a spare key.
/tinfoilhat
my objection has been noted and will be carried forward
I'm intrigued by the wooliness of 'going forward'. Theres an implication the the AI training has already been happening but you could also read it as we won't utilise your future posts and data but will continue to do so with older posts.
I feel like we're also missing a second opt out - theres two aspects of AI in social media. One is mining users data/post/images as content, but the other is serving that content to users. The latter bothers me more frankly and thats the opt out I'd like. In the same way as promoted and paid-for social media content is required to be flagged / declared I feel AI generated content needs to be declared also.
Opted out and have the confirmation email. Sending address looked dodgy as hell though 😆
Nice one, thanks for the write up, most useful.
Aye thanks for this Cougar, I've opted out.
I get closer and closer to ditching FB, but older family still use it.
In the case of Microsoft. Even if you pay and buy a licence, you are the product. And now they have spyware “but it is encrypted”…. err yeah with keys that they look after for you.
Or you could just, you know, not use that feature. It's all locally stored so under the users control anyway.
It's not like their competitors haven't had their own privacy issues, look at the recent are they/aren't they deleted photos issue on iOs.
The link from the blog worked for me fine- but not when I forwarded it to Mrs K. Strange
I've heard a couple people say they got stuck in a "you need to be logged in" loop. Logging into Facebook from an Incognito browser window fixed it for them. 🤷♂️
Your other half has an unusual name.
I opted out...thanks for the heads-up.
I'd delete my account if it wasn't for a few older people I needs to keep in touch with.
Got this email reply:
We’ve reviewed your request and will honor your objection. This means your request will be applied going forward.
If you want to learn more about generative AI, and our privacy work in this new space, please review the information we have in Privacy Center.
facebook.com/privacy/genai
This inbox cannot accept incoming messages. If you send us a reply, it won’t be received.
Thanks,
Privacy Operations
I fully subscribe to the principle that if you put any information or data onto a social media platform you should expect them to do stuff with it. If you don’t want it used, don’t give it to them.
Great point so long as no-one else ever puts your name or likeness onto social media.
I take the confirmation as meaning anything new added will not be used for training.
If they have already got a training data set then the chances of them even having the ability to remove an individuals data is remote.
I'm afraid I am still highly cynical about AI as I subscribe to the shite in = shite out school of thinking, and there is an awful lot of shite on the Internet
'kin ell. Did the opt-out but it took roughly 14 hours for the OTP to come through, thankfully they are instant now.
I’m very concerned about how AI is creeping into so many parts of our lives with little regard to anything other than the huge profits to be made.
Although as a SW Dev it has made a huge change to my productivity, quite incredible how much stuff it does for me now and how quickly I can get answers to stuff I need.
I occasionally press LinkedIn’s ‘rewrite with AI’ button
I have found some of the ‘AI’ grammar and spelling features of Word risible. All MS Office products have become noticeably worse in the last couple of years. I’d be unsurprised if MS were deploying the same models in multiple products. I’ll make no comment on the Apple spelling ‘AI’ that loves throwing apostrophes around like seasoning.
An AI writing aid I’ve been asked to try performs similarly to @cougar ‘s description. Short, simple text goes in and a mishmash of second language English and buzzwords comes out. Amusing, but I already see examples of folks using this kind of generated guff for real. 😞
our AI overlords might eventually triumph but I’m not sure we’ll understand what they want unless their language skills improve.
Did the objection thing. Response came back almost instantly. Less than 1min between the OTP and response emails.
They really do want to make it as difficult as possible, and make it look like you have to demonstrate why and how you have been harmed. It should just be a toggle in the settings, off by default (and the same goes for M$ etc.).
It's the AI models that don't have their own data source ecosystem that will have the biggest, potentially bankrupting, problems though.
I’m intrigued by the wooliness of ‘going forward’. Theres an implication the the AI training has already been happening
I found another form here which may be of interest.
Thanks Cougar.
It went through almost instantly,so I can only assume that the AI had no problem with my GDPR and cheese concerns.
Heh. AI processing requests not to use AI. How, ahem, very meta.
That second form is rather more insidious, you need to give reasons why you think your info has been accessed by AI.
Er, no ta, just delete it and stop sending it to third parties.
In honesty, I've not investigated it beyond "it exists" yet.
you need to give reasons why you think your info has been accessed by AI.
"Because you told me it had"?
Thanks for the link in your blog, have opted out.
However, as the form linked to asks for you email address. And it then emails you to confirm you've opted out. I can't help but think maybe I've been duped and the form is in itself an email gathering tool....
I don't think I was logged on on that form, so does it really link.back to me FB account at all? It looks genuine, but was on mobile so harder to check web addresses etc.
Not opened the confirmation email yet to check source address!
Uh, Facebook already has your email address? Obviously if you've entered a different one into the form from the address on your account then it's not going to work.
But, yes, it's genuine. You can get to it via the Facebook website/app itself, it's just buried in the menus.
I wrote "because I don't to" or something as the reason, which was immediately accepted.
Looks like multiple hurdles to discourage people from opting out - the notification without buttons on, form to fill in, verification PIN code by email you need to enter.
To be fair, without the PIN you could opt-out everyone you knew without their consent.
Only because of their clunky form that requires an email input rather than infering it from the logged in account.
I've tried multiple times now and still haven't received a PIN.
hmmm I'd forgotten about this, I objected after reading Alans blog & heard nothing... I finally got a reply today, that my objection had been honoured.. Yay only took 2 months
It was a weird one.
When I first opted out I heard nothing, days went by and I was reading posts from people saying they'd heard back straight away at the speed of automatic response. I went through the exact process exactly a second exact time and received an immediate reply.