At the Munich Security Conference last week, Facebook CEO Mark Zuckerberg stated that his company removes more than one million fake accounts every day. “The vast majority are not connected to state actors trying to interfere in elections, but certainly part of that is a state effort.”
Zuckerberg undoubtedly meant this fact to demonstrate the seriousness with which Facebook takes the threat of fake accounts, but fraud occurring at this level isn’t a sign of success but of a problem that is well out of hand.
According to Facebook’s latest Transparency Report, the rate of fake account creation is far worse than Zuckerberg reported. In the third quarter of 2019, Facebook says it disabled 1.7 billion fake accounts. Over the last four quarters, it disabled 6.6 billion fake accounts. To be clear, these numbers do not include accounts that were blocked from being created in the first place, which Facebook estimates to amount to “millions” more per day. For perspective, Facebook has 2.37 billion active monthly users.
Facebook continues to make it quick, easy, and free to create a new account and then use data analytics to identify fake ones after the fact. Facebook boasts that it identifies 99.7 percent of fake accounts before they are identified by users. The problem is that at this scale, the remaining 0.3 percent could amount to over twenty million accounts. And of course, there is no way of knowing how many millions of accounts avoid detection by both Facebook and vigilant users, though Facebook estimates that 5 percent of its accounts (118 million) are fake at any given time.
If Russian bots only need to change 40,000 votes in rust belt states in 2020, they will have plenty of fake accounts with which to carry out disinformation operations.
It’s time for Facebook to rapidly introduce validated identities on its platform. There are reasonable arguments that can be made for allowing people to remain anonymous or pseudonymous online. Those arguments do not apply to Facebook, which maintains a “real name” policy—no parody accounts, no avatars, only real people.
Facebook already has the process in place to validate identities, but it is only doing this for accounts that wish to purchase political ads. I went through the process myself and judge that it would scale well and make it much harder for trolls in bot farms to create fake accounts. (Bonus points: it uses the U.S. Postal Service to validate that a user lives in the United States).
Facebook should begin requiring identity validation for all new accounts immediately. New users could still join and use the platform immediately but should be marked as unverified and taken down within a short period if they do not complete the verification process.
Current users should be offered the opportunity to get verified accounts, receiving a verification mark when they do so. Within a short period, accounts that have not been verified should be marked as such for other users to see. Eventually, only validated identities should be allowed on the platform. Because time is short before the next election, Facebook could take the radical step of automatically blocking engagement in political discussion by anyone whose account is not validated. Doing so would provide a strong incentive for many people to get validated accounts.
Facebook will likely protest that it cannot be expected to take such actions. It is, after all, a business. That business, however, is selling ads to display to its user base. Instead of chasing the next billion users by making it far too easy to create an account (no credit card required, no captcha to fill out), Facebook could make the billions of users on its platform far more valuable by proving to advertisers that there are actual people behind those billions of accounts.