Every programmer knows the difficulties when it comes to updating a platform, monitoring user interaction or testing new functions. Live updates must work quickly and smoothly, and a user’s misbehavior must be stopped as quickly as possible. The failure of a site or overlooking an error would be the worst case scenario. But what if you run a platform with 2.6 billion users? Here the risk increases enormously. With every update, social media giant Facebook must therefore ask itself how they can keep the risk as low as possible so that none of their users suddenly find themselves sitting in front of an error screen or are neglected.
As they report in a recently published blog post, Facebook has now found a solution to this problem: artificial intelligence. The company’s AI team now simulates an identical copy of the platform before each update, full of bots and virtual users and tests each new function and each update first in this fake Facebook.
The AI developers of the social media company call the program Web-Enabled Simulation, WES for short. WES enables the AI-supported creation of a highly realistic, large-scale simulation of complex social networks.
“We’ve used WES to build WW, a simulated Facebook environment using the platform’s actual production code base. With WWW (the name is meant to show that this is a smaller version of the World Wide Web, or WWW), we can, for example, create realistic AI bots that seek to buy items that aren’t allowed on our platform, like guns or drugs. Because the bot is acting in the actual production version of Facebook, it can conduct searches, visit pages, send messages, and take other actions just as a real person might. Bots cannot interact with actual Facebook users, however, and their behavior cannot impact the experience of real users on the platform.”
The bots, which simulate Facebook users in WES, are represented by a specially trained AI model, which depicts human behaviour and interactions in the social network. You can add friends, share or link posts and interact in groups.
This handling of updates and new features is a secure solution to focus on the user experience and to detect possible errors early. Through the learning AI, data from the live platform can be used to predict every eventuality. This type of simulation is a model that is also worth considering for other platform operators to bring more security and efficiency to their maintenance work and the interaction between their users.
