Social Media's Misinformation Crackdown Falls Flat Ahead Of Midterms

Social Media's Misinformation Crackdown Falls Flat Ahead Of Midterms
Even with new rules and multimillion dollar "war rooms," social media companies are finding it difficult to keep disinformation off their platforms.

After the 2016 election, social media giants took steps to weed misinformation out of their platforms. But even with "war rooms," hundreds of millions of dollars and hours of testimony on Capitol Hill ahead of the midterm elections, it doesn't look like anyone's solved the problem.

Facebook, Google and Twitter all worked to cut out fake news, stop misinformation or rebuild public trust in their own platforms. Facebook created a special "War Room" staffed with engineers and other experts who identify and stop the spread of fake news. Google announced a $300 million "Disinfo Lab" with goals much like Facebook's. Twitter updated its rules so accounts could be banned for new infractions, like using a stolen avatar photo or profile biography.

To some extent, these plans seem to be working. Each outlet says it successfully shut down accounts associated with spreading misinformation during the midterms. But experts found these companies still frequently failed to catch problems their tools promised to stop.

Google keeps running ads from groups that publish polarizing misinformation. In September, the Campaign for Accountability reported it was able to pose in Google's system as the Internet Research Agency — the same Russian troll operation the feds indicted for interfering in the 2016 election. Google allowed it to pay for and publish ads that appeared on websites like CNN and CBS This Morning.

Facebook's new advertising tool promised to flag potentially polarizing political ads for further review, but it doesn't seem too be working correctly. A USA Today analysis showed Facebook consistently removed non-political advertisements that were either written in Spanish or mentioned the words "african-american," "latino," "women" or "LGBT."

Twitter also reportedly failed to remove accounts that were publishing fake and conspiracy news even before the 2016 election. A Knight Foundation study found of the 700,000 Twitter accounts linked to fake news or conspiracy outlets during the 2016 election, more than 80% were still running at the start of October 2018.