Key-words: fake account - social media companies - face recognition The largest social media platform in the world (2.4 billion users as at September2019) is testing users to authenticate their accounts through face recognition so they can continue to use their accounts. If “selfie recognition” testing, as Facebook calls it, is considered successful, the social network will use this authentication method to prevent the proliferation of fake accounts on the platform. In addition, the company is preparing “stronger verification requests” to ensure that users can continue to use their accounts. As a guarantee of privacy for users, the system being tested is called “face detection” and it isn’t used the term “facial recognition” because it reiterates that they only use the system to find out if the other on the side of the account there is a human being or not, not the human being who claims to be the account. The circumstance was confirmed during a meeting with journalists at Facebook's London headquarters in which the Portuguese “Observador” participated. According to a high-level manager of the Platform Community Integrity team, this "self-capture" system serves to "prove that the person behind the account has good intentions. (..) One thing that helps us is knowing that someone is behind the camera.” The manager says this mechanism is only in the testing phase and "for verification only": the system is activated only at times when Facebook considers that the account was created for harmful purposes and is not controlled by a real person. That is when it is a machine or algorithm that manages the fake account. Regarding user privacy, the manager promises that the images collected are not being compared with other images and this method "only serves to confirm that there is a real person behind the account". He also notes that if this method is no longer being tested, it will deter people who create multiple social media accounts to mislead other users. About this new mechanism, that works very similarly to facial recognition in the latest smartphone models, it has been specified “there are several situations in which we require people to provide us with more checks. By way of example, the most obvious are political ads. If you want to have a political ad on Facebook you have to verify your identity and location. Until you do that, you can't post political ads. This is to ensure that who makes these ads is who they say they are. ” If Facebook considers the account to be fake, before being able to log in again, the user must allow access to the front camera of the device they are using. Then, while the camera is shooting the user, several steps follow - such as “look right or look up” - to confirm that it is a human being operating the account. Regarding the privacy issues that this tool can create - as you may not want Facebook to reuse your movie on your account - the social network no longer feeds information. "We are not sure when and how we will require people to give further proof of their identity." However, there is a promise: the goal is to keep the community safe and minimize bad experiences. “If there was a magical way to verify people's identity and authenticity while minimizing the biased data we collect, that would be perfect, ” the manager says. The social network also says it is working with “organizations and partners” to improve and facilitate authentication technology: as an example, in 2017, Facebook tested self-recognition that did not advance. Instead of filming the user, the platform would ask the person to upload a profile photo. Being a fallible method, and having created controversy, it did not advance.
Source: Observador.pt , Eleito melhor jornal generalista 2018.
To read the PDF click here. |