The 10-year social media challenge raises the issue of facial recognition

From the Agency of Brazil:

If you are a social network user Facebook, you can hardly see the publication of a friend with two photos, one of the 2009 and another in 2019. The game, nicknamed the "10-year challenge" (or # 10 year old, in the hashtag ), has since become a major "joke" . "

The challenge has won all forms of adaptation, from people posting photos comparing their images over the last ten years to a hook for the comparison of artists, politicians, places and situations. has become a resource for discussions ranging from the personal evolution of users to political disagreements on the site.

However, the popularity of the challenge has also sparked debate from experts in computer security and data protection Kate O'Neill, American consultant and author of books on digital technologies, interviewed social networks and the world's largest technology magazine, Wired highlighting how published images could not be published used to "train" the Facebook system that performs facial recognition of users.

Automatic Marking [19659007] When an image is published, the social network performs this identification. This feature appears, for example, when she "suggests" tagging the user or her friends on photos.

However, the platform began to perform these "badyzes" on all images and to "warn" the person when publishing a photo. The argument was that the user was more in control of the content on the site. However, the deployment of this feature has been criticized. Facebook reacted and configured the feature as an option that can be disabled by the user.

Recognition, even with this tuning device, is performed via a technology that "learns" how to improve this procedure as it is received. more data or more photos. Hence the questioning of Oneil and other experts on how such 10-year-old images could "feed" the Facebook database and "train" their systems.

"It can happen, it's a very objective metric that systems have to learn, the big challenge is to create universal standards, it's an objective standard because it has set years." used to train these systems, "evaluates the researcher in privacy and professor at the consulting firm DataPrivacy Brazil Renato Leite.

Facebook

Faced with the repercussions of doubts, Facebook has positioned itself on the subject. is a meme created by users of our platforms and spontaneously activated. Facebook has not yet launched this wave and the meme usually uses photos already present on Facebook. We have not won anything at this meme (apart from reminding us of the dubious fashion trends of 2009). As a reminder, it is useful to specify that Facebook users can choose at any time to leave facial recognition active or not, "says the company in a note.

However, Leite indicates that the photos posted by users, by date 10 years ago) were not necessarily available or had been published on the social network. "If the photos are published on the timeline and that Facebook performs the badysis because this is part of the Algorithm, they will benefit. Although Facebook says it's a third-party initiative, the key question is whether the company actually uses the photos to form its algorithms with these metrics. "

This doubt, adds the expert, raises a reflection on the need for greater transparency on the part of the company vis-à-vis its users so that" they know how their images are used and for what purposes, adding that the impact exceeds the users themselves, because the systems even map out to people outside the social network when a photo is published, thereby generating a unique identification or a "ghost profile."

Creating a Discussion

A Veridiana Alimonti, policy badyst for the digital rights organization Eletronic Frontier Foundation, questions the fact that, Regardless of whether society benefits from gambling or not, the controversy surrounding the challenge has been beneficial in challenging facial recognition.Its entity, the EFF, has conducted studies on threats such as the malicious use of biometric data s, sharing without transparency between public authorities and private companies or at the expense of people resulting from a false identification of their identity.

According to Veridiana, such debates can help citizens become aware of how they disseminate their information on social networks. like Facebook. "If this discussion draws attention to that, let us warn about how people treat platforms in relation to the information they have and what uses they make," he said.