Facebook used machine learning to remove 6.6 billion fake accounts

The AI analyzes active accounts and also studies profiles’ behaviour and interactions

Facebook has implemented a machine-learning tool that has already removed 6.6 billion fake accounts in the last year.

The social media giant told ZDNet that this number doesn’t include the millions of attempts that are blocked daily when attackers try to create fake accounts.

The platform has started using a technology called Deep Entity Classification (DEC). It uses machine learning to analyze active accounts and also studies profiles’ behaviour and interactions.

It analyzes the different profiles and groups that people interact with and also looks at how many friend requests people send and how many have been accepted.

DEC is able to map out every profile on Facebook through its friends, groups, pages and more. This allows the system to see what types of accounts appear to be faking their identity.

“So you can see, through a number of signals, if a user is trying to misrepresent their identity. It’s not so much about the content of an account, but about how that account interacts with others on the platform,” Daniel Bernhardt, the engineering manager at Facebook, told ZDNet.

Facebook says this new tool ensures that attackers are unable to create fake accounts as easily as they used to before.

Source: ZDNet