Facebook Logo
Electronic cables are silhouetted next to the logo of Facebook in this September 23, 2014 illustration photo in Sarajevo. Reuters/Dado Ruvic

The social media giant is developing an advanced artificial intelligence software that could determine what is happening in the photo beyond just simple face detection. In cases where in the user is intoxicated, the software will have the ability to warn against uploading a laughable and witless selfie.

According to Yann Lecun, head of the Facebook Artificial Intelligence Research (FAIR) lab, his team believes that the future Facebook AI would act like an assistant to its users. The AI also aims to stop crazy behaviour or at least warn users when they are about to initiate something they might regret. LeCun also wants to protect users from their friends with the help of the advanced assistant, which could warn users that another user is about to upload a potentially embarrassing photo without his permission.

To illustrate, FAIR's lab head pictured a Facebook that has the ability to prevent users from posting unflattering pictures. Aside from identifying the user's face, the social networking leader would also note the time when the user is posting, it would also identify whether there are drinks or even unfamiliar people involved in the picture and, most importantly, if the photo is appropriate for the user's school or workplace. With the given factors, the AI would then be able to draw a conclusion and thus provide an appropriate warning for the user.

In addition, the tool is mainly to build image recognition technology that has the ability to differentiate the user's drunken self and sober self by using an artificial intelligence called "deep learning," which is a technology that was spearheaded by LeCun and other academics. At present, Facebook has the ability to identify the user's face and his friends' faces in photos that are being posted in the social network. It also allows the user to easily tag other users in the photo with the right names, Wired reports.

Based on Gizmodo's report, some users do not mind uploading drunken selfies. Some users might not even like the thought of having some Facebook's robots to censor the contents of their posts. While the idea of the "deep learning" technology is most exciting, however, some still find the concept a little off.