When Russian memes overtook actual political ads during the 2016 election season, Facebook knew it had a problem. But its solution isn't making things any better.
At the end of May, Facebook unveiled its automated system that flags political ads and requires their makers to authenticate their identities. The system also labels who created an ad and stores that information in a searchable database.
But ProPublica has been monitoring Facebook political ads since September, and what it's found hasn't exactly been uplifting. In just a few weeks, ProPublica has documented several cases where non-political ads were marked as "electoral" or "issue" messages. Many of those misdirected flags hit news stories on one of the broad national issues Facebook is sweeping ads for, including crime, health, and education. Meanwhile, ads from Democratic senators and advocacy groups on both sides of the aisle haven't been scooped up.
Many news organizations that have fallen victim to Facebook's tracker have opted not to run ads until it's fixed, ProPublica reports. The New York Times' CEO even publicly detailed his concern that Facebook is getting news all wrong.
Facebook's defense? “Enforcement is never perfect at launch,” the company told ProPublica. Read more about the Facebook project at ProPublica.