Facebook has created a bot system to simulate users in order to find glitches and security flaws that its human researchers may not be able to catch.
In other words, Facebook has essentially created a parallel social network with bots that are able to post and like content and even spread spam. The social media giant is using this network to try to find ways to stop people from abusing its platform.
Facebook has explained its ‘Web Enabled Simulation’ a recent paper that shows that its bot system is filled with fake users that are based on actual users’ behaviours.
For instance, some bots on the simulated network are designed to seek out content that goes against Facebook’s rules, while others are trained to try to invade fake users’ privacy.
Although software simulations are not uncommon, Facebook says that it takes its system a step further because its bots operate in something similar to a real social media platform. The bots are able to send friend requests and perform other functions that human users do.
Facebook uses the system to test new features for a future update. For instance, if it launches a potential feature in the simulation, and the bots are able to exploit it, Facebook can study the flaws and fix the problem before releasing the update to actual users.
The social media has outlined that there won’t be interactions between the bots and real users because they are isolated from them.