After a year of negotiations, Facebook, Microsoft, Twitter and YouTube agreed to the European Union’s hate speech code of conduct , which commits them to review and remove the majority of valid complaints about illegal content within 24 hours and to be audited by European regulators. The first audit , in December, found that the companies were only reviewing 40 percent of hate speech within 24 hours, and only removing 28 percent of it. Since then, the tech companies have shortened their response times to reports of hate speech and increased the amount of content they are deleting, prompting criticism from free-speech advocates that too much is being censored.
The new team of human raters, which Facebook calls the “feed quality panel,” are key to surfacing this meaningful content. Each day a typical panelist rates 60 stories that would actually appear in their own News Feeds on a 1 to 5 scale, judging how interesting they found the posts. They also reorder their own feeds to show how their priorities differ from the algorithm’s, providing what Facebook calls a transposition score. And they write paragraph-long explanations for why they like or dislike certain posts, which are often reviewed in the News Feed engineers’ weekly meetings. Facebook also regularly conducts one-off online surveys about News Feed satisfaction and brings in average users off the street to demo new features in its usability labs.