During her hearing before the United States Senate Select Committee on Intelligence, Sheryl Sandberg, Facebook's current Chief Operating Officer, has stated that during a period of six months the social network had to delete 1.27 billion fake accounts.
According to Sandberg's statement, the operation was driven by manual review methods, as well as with the help of automated review methods powered by machine learning, computer vision, and artificial intelligence.
The process of keeping the fake accounts number in control allows Facebook to discover malevolent actors which would take advantage of the platform to possibly spread fake news and skew the public opinion to match their own agenda.
"We have more than doubled the number of people working on safety and security and now have over 20,000. We review reports in over 50 languages, 24 hours a day," said Sandberg. "Better machine learning technology and artificial intelligence have also enabled us to be much more proactive in identifying abuse."
Artificial intelligence and machine learning used to protect the social network's users
During the hearing before the Select Committee on Intelligence, the Facebook representative also mentioned that the social network also works on removing spam content from their platform, with an astonishing number of 836 million pieces of such content being removed during 2018's first quarter alone.
Account networks designed to mislead users and misinform them are also the target of Facebook's automated and manual review processes, this category of threat possibly including both real and fake accounts.
Also worth noting is that, as Sandberg stated during her hearing, "Facebook has a security team dedicated to understanding how bad actors attack individuals and networks, building defenses against such attacks, and reacting quickly to mitigate potential damage."
With the help of this team of security experts, Facebook wants to stop the spread of malicious content to their users, send automated custom alerts to people affected by various types of cyberattacks, as well as making sure that accounts are not being hacked in using artificial intelligence-powered techniques.