Facebook is changing the way its Trending Topics works by updating the company’s guidelines and workflow.
The social network faced accusations earlier this month of political bias in its Trending Topics section. The allegations came after an investigation by Gizmodo into Facebook’s practices and methodologies for curating news, where editors reportedly routinely suppressed conservative news.
After an internal investigation, Facebook claims it found "no evidence of systematic political bias in the selection or prominence of stories included in the Trending Topics feature." Facebook says it approves equal amounts of conservative and liberal news.
Making changes anyway
However, Facebook also admits its "investigation could not fully exclude the possibility of isolated improper actions or unintentional bias in the implementation of our guidelines or policies," which is pretty obvious whenever you have human curators.
Facebook is taking steps to ensure political neutrality in the Trending Topics section moving forward.
Humans will remain a part of the Trending Topics equation, but employees will be retrained on updated guidelines with "more clear" terminology. The company also promises to put in place "additional controls and oversight around the review team."
How news is surfaced is also getting a slight tweak. Facebook will no longer use the "Media 1K" list, which was a list of external websites the company used as reliable sources to surface stories that weren’t currently trending. The list was used in conjunction with an algorithm to find breaking news.
These changes were implemented after Facebook’s meeting with leading conservatives last week. The social network promises it will continue to be "a platform that is open and welcoming to all groups and individuals.
While Facebook’s News Feed is entirely algorithm driven, its Trending Topics section still requires human curators to ensure the validity of stories, as well as surfacing important stories that aren’t trending. Some have argued Trending Topics should only be picked by algorithm, but others believe that human-created algorithms are inevitably biased in the first place.