That’s a cool idea. Needs to be a browser extension so that it is automatic, or even in-built browser behaviour, such is the scandal of the modern world.
However there’s a 100% foolproof technique here … don’t post photos online, and in particular not where it will be accessible to prying government or Big Tech. But, as the article says,
In many cases, we do not control all the images of ourselves online; some could be posted from a public source or posted by our friends
Not posting photos online is still 100% foolproof in the sense that even if you use cloaking software, your “friends” who post photos of you probably aren’t using cloaking software, and no software can defeat anything if the software is not used.
But sadly, as the article observes, if your “friends” keep on posting uncloaked photos of you, the only way of combatting that is by winning out in the numbers game i.e. posting even more cloaked photos of yourself.
As an option, that’s a fair suggestion but I actually own a real digital camera, as well as spyphones. I aint going to be hacking the firmware on my digital camera to build in that functionality.
Also, and in any case, it wasn’t clear to me whether the cloaking software itself is capable of recognising what a face is so that it only cloaks faces where it finds them in an image or it just cloaks the whole image. I wouldn’t necessarily want it cloaking images that don’t contain faces. So I think it may need to be an option, wherever the functionality is offered.
Also, from the perspective of purity, I would prefer to own an original unmodified image. I understand of course that an image is already “modified” heavily before I even get my hands on it. I have a fairly large number of photos in my collection and I would prefer to put such hackery at the boundary between “private” and “public”.
As such therefore, I would need it in the mail client as well as the web browser - which I guess highlights the trade-offs between what you are suggesting and what I am suggesting.
It wasn’t clear to me what the effect would be of cloaking twice. So if I do have the functionality built-in to the Librem 5 and I download from phone to my collection and then I upload via browser / send via mail client and they cloak again, is that a problem?
Do we need a new JPEG metadata tag to indicate cloaking status? Would you trust the value of the tag if such a tag were defined? (That wouldn’t be a problem for me as my media workflow is mostly WadeSoft so I could easily strip the hypothetical tag unless the image source is Librem 5.)
I still don’t think they are pointles and the integration idea is good, tagging included. I’d also include here an authentication proof method for the other side of the problem: create hash as proof of the image that gets saved and maybe also uploaded somewhere you want for later comparison.
You would ideally strip the tag completely as the image leaves the private domain. The point is that if multiple internal components might add cloaking, there may be benefit in explicitly keeping track of whether cloaking has been applied so that it doesn’t get applied twice.
There was a test of face socks, a stretchy mask with eye holes that have a inkjet printed face on it. The test had acquired some President Obama face socks and the tech did not flag it as masking and tried to pull an ID as a real face. To a human eye though they are terrifying to see, hyper uncanny valley to wear on the street.