| Hey all, it's Sarah Frier. On Wednesday, Republican senators laid into the chief executive officers of Facebook Inc., Twitter Inc. and Google for anti-conservative bias. They asked pointed questions like: How many of the employees and content moderators are liberal versus conservative? And how many high-profile posts from Democrats have they removed, versus posts from Republicans? (They don't keep data on either question, the CEOs said.) Senate Democrats criticized their colleagues for bullying the companies, saying the hearing was a transparent effort to pressure the online platforms to make calls favorable to President Donald Trump days before the critical election. They had a point. However, for months, the Democrats have been pressuring companies, too. They've repeatedly said the platforms aren't doing enough to correct misinformation about voting. Whatever happens during the U.S. election next week, the political party that loses the presidency is likely to lay the blame at least partly at the feet of the country's social media giants—either for doing too much or not doing enough. That means scenes like Wednesday's hearing battering the tech CEOs will remain a favored political pastime for the foreseeable future. This state of our politics is, in part, the companies' fault. For years, executives have argued that their websites and apps were neutral platforms. But they always had an algorithmic bias toward content that sparked an emotional response, spreading those posts farther and faster. That amplified more extreme views and exacerbated arguments and division online. Now that the companies are finally reckoning with their societal impact, the fixes can look superficial and even random. That's in large part because the companies haven't fundamentally altered their algorithms—they're reacting to and moderating content after it's already up. There's plenty to criticize about social media content moderation—it's often applied inconsistently, and even unfairly and untransparently. But the reason for that isn't a conspiracy to silence certain users. (Indeed, claims of anti-conservative bias tend to go viral on these same platforms.) Rather, the culprit is the companies' own focus on getting bigger and driving engagement, rather than cleaning up messes. Eventually new artificial intelligence systems could effectively tamp down on hate speech and viral misinformation. But right now, the computers aren't smart enough to understand all the contours of human discourse. So Facebook, Google's YouTube and Twitter are compensating by building election resource hubs, where they'll give verified information. And they're adopting dramatic policies—like no political ads at all on Twitter, and a link to verified information on every Facebook post about voting—so that companies don't have to make decisions one-by-one on billions of posts. These changes have come piecemeal, over the course of hundreds of incremental announcements, because of consistent pressure from regulators and other third-party critics. The companies are so powerful, hosting so much of our election-related discourse, that tiny tweaks to their policies can have massive ripple effects. Wednesday's hearing showed that the politicians have learned this pattern, and know the fastest way to get the companies' attention to a problem: a loud critique.—Sarah Frier | 
Post a Comment