Facebook will soon use your photos, posts and other info to train its AI
thejournal.ie
thejournal.ie
It is absolutely wild being in the "before times" (i.e. before any real legislation is applied to the use of data for AI training) and seeing statements that basically amount to "eh if we feel like it, we'll exempt you".
Not very often you see an "if" in statements like this.
https://soatok.blog/2024/02/27/the-tech-industry-doesnt-unde...
>In the case of ImageNet, the current state of the art accuracy is obtained by Facebook AI’s weakly supervised ResNeXt-101-32x48 model. It delivers 85.4 percent top-1 accuracy, outperforming by a significant margin the recently published EfficientNet network. . The model is pretrained on one billion public Instagram images, which contain 1,500 ImageNet-related hashtags.
Instagram locked by account for posts that I didn't make (I don't post on IG, nor did anyone else post on my account) and gave me 90 days to appeal the decision. As I had 2FA turned on every recovery method involved an EMail (not my idea). So I could not appeal and they locked my account forever. To me not a great loss.
Is it an image that is searchable on the web ? or A pic visible by a few outside of your FB network ? A pic visible to anyone besides the owner ?
This seems relevant since they have trained on this basis
The whole idea that you can dictate how an image you shared online without terms, is used, is hilarious.
"This Instagram image is mine, no one can share it without my permission!" always makes me chuckle.
They haven’t found a way start doing it already?
I'm sure they'll claim that the friend has the rights to the image and the friend gave permission via a click-wrap contract of adhesion.
I'm sure it's in the TL;DR TOS in the US already accepted long ago in simply continuing to use facebook services. Zuck said it best when he called facebook users dumb fscks himself.
This company's (and really many of its peers') sociopathic behavior is right out of a late 80s sci-fi thriller.
We and our 71 vendors process data for the following purposes:
Use precise geolocation data. Actively scan device characteristics for identification. Store and/or access information on a device. Personalised advertising and content, advertising and content measurement, audience research and services development.
I would bet good money that they still have everything I ever posted in their databases (with a "deleted" flag set), and that there is absolutely no way for me to reject consent for them to use all of that to train their bullshit engines.
Their comments about using public images and commentary being massive dataset to leverage came across as worrisome but expected.
--
To: Meta / Facebook
Re: AI training
I am a user of your platform in <Country>.
I’ve recently been given to understand that you plan to start utilizing my personal data (photos, posts, etc) to train your AI models. I am writing to make it clear that I DO NOT consent to this use of my information.
I tried to find a setting on your website to “opt out” and have been unsuccessful.
Be advised if you utilize any of my data for training without my consent, I will litigate to the full extent of the law, including seeking an order to compel the destruction of any resulting models, weights, data, source code, etc. that has been generated from it. Should such an order be granted it could be a very expensive consequence for Meta given the likely size of the resulting dataset that would have been “tainted” with inputs from my personal data.
Further, I strongly advise you make this an “opt in” process for users – in addition to mitigating your legal risk, it’s simply the right thing to do. Expecting that you already somehow have permission for such an invasive use of your users’ data and constructing the feature as “opt out” is, frankly, contemptible.
I'd love to delete my account, unfortunately some groups I'm a member of continue to insist on using the platform for carrying out business. :-(
Your threat has about as much weight as a child hitting a supertanker.
Do you really think Meta is going to stop collecting data based on your advice ?
2) I understand your reaction, and that's an amusing analogy, but it's presumptive to assume I lack means to follow through.
3) I agree one email is a drop in the bucket. If you're similarly appalled I encourage you to act on your principal and signal it's unacceptable, in whatever way you choose. Even sending a tweet is better than apathy.
Large tech companies make transparency 180's all the time after realizing too late they hatched a PR turd. Enough of an uproar over a variety of channels could conceivably have some effect. I agree they're unlikely to abandon training altogether, and I'm not advocating that, but a clear and working Opt-Out button would be a good start.
the right to request their records, subject to Privacy Act exemptions;
the right to request a change to their records that are not accurate, relevant, timely or complete;
and the right to be protected against unwarranted invasion of their privacy resulting from the collection, maintenance, use, and disclosure of their personal information. https://osc.gov/Pages/Privacy-Act.aspx
I say seems, because even though they aren't a federal agency, they should be held to the same standard at a minimum. It's our data, they don't own it, and they don't have a right to injure us for it.
Exactly. So Facebook is not violating the Privacy Act and your comment is irrelevant.
When you voluntarily chose to use their service you agreed to give them your data. It's now their data to do with what they like. If you don't like it don't use their service. Pretty simple concept that hasn't changed in the last 20 years.
That is the problem right there.
"It seems you just committed criminal perjury! I say seems, because even though you aren't lying under oath in a court of law, you should be held to the same standard at a minimum."
No, dangit, "seems" is not some kind of magic word to bridge objective reality with a daydream.