3!
That's two things I learned today, make that three!
One, "Ever"???
Two, a "Danny Baker" exists
Three, apparently there's a "Radio 5"
Millions of images stored by Ever, a photo album app, are being used to sneakily train facial recognition systems – and the only way you'll find out is if you read through the 2,566-word privacy policy. The policy for Ever states that: “To organize your Files and to enable you to share them with the right people, Ever uses …
Well I am so reassured.
FaceBook is also in compliance with all applicable laws, look how well that train wreck of a privacy violator is going.
Ever is clearly a project masterminded by some seriously smart people. Ever wants to make its money from facial recognition, so it creates a photo-storage space and suckers people in with a free app to create a database it can use to train its statistical analysis machine. That is smart, no doubt there.
And, since it's in the TOS, they have no risk of a lawsuit. And of course, people have flocked to the thing like the sheep they are.
<sigh>
Quote
To organize your Files and to enable you to share them with the right people
Stop right there. That is an immediate red flag. sorry No, No and once again No!
I do my own organizing and if in some far off universe I want to share my pictures then I will do it myself.
Numpties if they think that anyone with half a brain would want to use this service. Oh wait... How many people voted for TurnipHead in 2016?
There's probably a line in the TOS about changing the TOS as they see fit. They probably even told everyone they'd changed them e.g. "Please read and accept our new TOS!"... What do you mean which bit changed? that's for us to know and you to meticulously compare the two versions of our TOS that you studied in great detail before accepting...
I can't tell from the article how well the data would be anonymised in training AI.
But I spot a possible analogy to data we used to train and evaluate computer speech recognition systems, when I was doing post-doc in the subject. The big speech sample datasets we used came from third parties (the most standardised was compiled by TI and MIT), and were designed specifically for that purpose. There was no data that we'd ever thought of calling "personal" on there. The maximum possible extent of our interest in any speaker was "what's that accent, mannerism, or speech impediment, and why does it cause the algorithm to do XYZ". But if we had wanted to ask, that information would not have been traceable.
That was back in the early 1990s. Ideas of privacy and personal data have obviously changed since then.
Trouble is, what to do about other people uploading photos of you? If you voice your concerns to them they will turn against you, slash your tires/tyres, disbar you, disown you and disinvite you from all future gatherings.
A similar problem arises when you, say, sign up your children for the local soccer/football team. The team is likely to only use Facebook for communication, and if you inform them that you avoid said data-slurp they give you a red card.
Then there's the Dick's Sporting Goods mobile app. Free, of course. As in "all your data are belong to us".
Some nuggets from the 'Privacy' Agreement:
- We use cameras in our stores for security, operational and marketing purposes. This technology may capture demographic and personal information about you.
- if you choose to access, link to, or log-in to our Websites or Internet Marketing Channels or otherwise communicate with us through a third-party social media service, you are granting us permission to access and use the information that you post or store on the applicable social media service in accordance with the privacy policy of that service and the privacy settings that are applicable to your account, and to store the user name and password you use to log-in to the applicable social media service
- We may combine certain Personal Information and non-personal information collected online and offline, including information collected from third parties;
In short, they collect as much information as possible, by whatever means we can think of. Buried farther down I found:
"[We may use your information for] Evaluating your application for employment."
BOOM
There seems scope for a little innocent amusement here. Uploading many copies of the same photograph - or even better, multiple copies of several different photographs of the same face, all tagged with different names and many photographs of different people all tagged with the same name. Maybe for the latter, the output of one of these systems that are supposed to generate photo-real images of faces.
There's only 1 rule for personal information: If you upload it to the internet it's not personal anymore.
It will be monetized if possible.
It will be compromised at some point in the future.
It will be used for purposes you never envisaged.
It will be no longer available at some point in the future.