As the 10 Years Challenge is going viral on Facebook, Instagram, and other social networks, the AI algorithms of these platforms will be the winners.
It’s the “meme” of the moment. If you use social media, especially Facebook, you will not be able to escape it: you have probably noticed since the beginning of the week that many members are displaying photos of their profile from 10 years ago compared to their current face. If this “challenge”, whose origin is unknown, may seem as amusing as it is innocuous, it could delight developers of artificial intelligence algorithms based on facial recognition. In any case, this is the point of view developed by an author, who explains it in a very long post in Wired, following a tweet that triggered many reactions.
Kate O’Neill also explained that for profile photos on Facebook, the date the photo was posted would not necessarily correspond to the date the photo was taken. Thanks to the same Facebook, most people have added this context (for example “me in 2008 and me in 2018”), as well as other information, in many cases, about the location of the photo and how it was taken. In other words, thanks to this challenge, algorithms now have a very large amount of photo data of people from about 10 years ago and the same people today.
In this regard, the author also recalls that recent years have been rich in examples of social games and similar games designed specifically to extract and collect data, the example coming to mind being the Cambridge Analytica scandal, with the massive data extraction of more than 70 million American Facebook users.
So of course, one can wonder who benefits from the crime, and what relevant data emerges from this type of data. Because the analysis of the photos posted as part of Facebook’s #10yearschallenge could still cause some problems for artificial intelligence and make them sweat a little bit about atypical cases in which Internet users enjoy publishing just about anything. But it is also the role of machine learning, and even its essence: to forge its competence on the data, and ultimately to be able to sort it out. In this context, it is conceivable that people who post pictures of their hamster also participate in strengthening the algorithms.