Linus Larsson: Your picture from 10 years ago is mumma for datagrävarna News

It appeared as suddenly as nättrender usually. Suddenly, everyone would load up two pictures of himself on social media, one from ten years ago and a current.

to show that one does not age so badly. To bid on that you do it. A tag and so it started: the Ten year challenge, tioårsutmaningen.

basically, it is the right of the innocent, of course. But now we live in the times we do and it is impossible not to think about the back of to hand over data on the network. The one that did it was the author, and lecturer Kate O'neill.

https://twitter.com/kateo/status/1084199700427927553

in an article in the Wired and her points hold: the Information tioårsbilderna is exactly what is needed to train system for face recognition, such a technique as a bit sloppy tend to be grouped under the umbrella of artificial intelligence. Or, more specifically, machine learning.

A prerequisite for ansiktsigenkänningsteknik, or what maskinlärning that time, large amounts of well organized information. It is by looking at the ten thousand photos of kittens a software can learn to recognize just kittens. In the same way it is by looking at the ten thousand (and preferably more) facial images with ten years between which the can work out their algorithm, how the ageing looks. Then can look at a picture and make a reasonable guess about how the person will look like ten years later.

There are of course both good and bad. Technology can put dangerous criminals. It is said that cases of missing children in India have been identified by digital face recognition. At the same time opens the way for extreme surveillance. The cell phone may turn off, you can avoid the internet. But you can't turn off his face. Just look at China, where the technology is built into the rating system that will reward the obedient citizens and punish those who do not comply.

often already there on Facebook with the dates and everything. But as Kate O'neill writes, so it is seldom exactly dated and defined as in tioårsutmaningen. Another criticism she received is that people put up tramsbilder, that it dirties the data and make it difficult for Facebook's and Google's giant datakörningar to train their algorithms. Like this:

https://twitter.com/JohanHilton/status/1085499209263054848

in it, but Kate O'neill is also right when she says that the software can easily identify the real faces and clear away the rest.

However, I think another human driver can make it for the AI emperors plans. The more 2019-photos sminkdolda wrinkles, belysningsmaskerade bags under the eyes and bortvinklade dubbelhakor, the more distorted the picture will get the algorithm of what it means to be older.

We have GDPR, Cambridge Analytica-disclosures and enforcement cases against Facebook. But in the case tioårsutmaningen it is as soon as the vanity that saves the privacy of the individual.