[ad_1]
Photos: Kim White / AFP and Ludovic Marin / AFP / Editing: Alessandro Feitosa Jr.
You may have seen posts on social networks with the tag # 10yearchallenge. Maybe you even participated. The joke is to publish a photo of you from 2009 and one of this year so that everyone can see how much you have changed in ten years of gap. But what looks like a harmless meme can constitute a vast source of information for companies in regards to artificial intelligence training, as author Kate O'Neill argues in an article published in Wired.
• "Do not Track": the tool
O Neill, technology expert, says the "10-year challenge" could be used by someone who wants to train an artificial intelligence to facial recognition and identification of age. Think: using a simple search for Twitter tags on Facebook, Facebook or Instagram, everyone can have access to a set of data containing many photos, almost all having the same person at two different times, separated by ten. Years C is a full-fledged dish for those with this intention: organized and identified data, ready to be badyzed by algorithms to identify transformation patterns in ten years.
The question is: does it really change something? We have almost all accounts on social networks since pretty much that time. My Facebook account, for example, was created in 2007 and does not contain any photos of that time as I cleaned up a few years ago. It may be much more common in other countries – here everyone was on Orkut in 2009, and Facebook was not a big thing for us that year.
It is therefore not impossible to get pictures of 10 years with this already posted on social networks. O & # 39; Neill argues that using this method, there may be a lot of noise between the data, but it is debatable – the photos have EXIF data, which indicate various information about the data. where, and that seems to be the main problem here, when they were taken. Even if there are photos scanned or posted on wrong dates, it is possible to obtain a very reliable mbad of data only using what is available.
Recognition and badysis of facial data may have beneficial uses, such as looking for missing children, or simply of a commercial nature. , such as data collection for advertisers of the São Paulo metro. But there are also the strangest and most harmful ones, like China's bizarre social credit system or the biased and imprecise system used by the American police.
Posting a photo of you in 2009 with a photo of 2019 may not have had a significant impact on your own privacy, for the information and data you have already shared up to present. Today, it's virtually impossible not to be tracked – you have to stop using all social networks, call VPNs, avoid smartphones, among other things quite extreme, if not impractical, lifestyle measures that we know today. 19659004] Whatever it is, we must think before publishing anything. As complicated as it may seem, it is precisely the data that we produce without considering the fact that they function as a fuel for this type of technology. And we already have concrete examples that a few years ago seemed to come from sci-fi movies or books presenting dystopias of the future. The Cambridge Analytica scandal proved that it was possible to do a lot using only stupid quizzes.
Even with so many follow-ups, we always have some control over what we share or not, and we turn away from it without thinking, just to participate in something that everyone does, seems rather naïve. An apparently harmless (and individually) meme, used on a large scale, may well pave the way for further exploration of our data.
Source link
