A leaked internal memo from Facebook indicates the world’s biggest social network will rely on human editors to pick stories highlighted in its upcoming “Top News” section.
The editors will drill into the sourcing of articles to determine their originality and reliability. They will prioritize stories broken by local news outlets, according to a report by The Information, which obtained the Facebook memo.
Facebook is testing a version of the tab that shows stories from The Wall Street Journal, ABC News, CBS News, National Geographic, BBC, HuffPost and The Hill.
That doesn’t necessarily mean these publishers have struck deals with Facebook to redistribute their stories. A report by the WSJ last month said the company was willing to write checks for as much as $3 million a year to license stories, but it’s not clear which publishers have struck deals with Facebook.
NBC media reporter Dylan Byers last week said small-to-midsize publishers were more enthusiastic than bigger media organizations about working with Facebook.
Facebook’s editorial procedures for selecting stories are certain to raise questions about news judgment. They likely won’t immunize the company from accusations of anti-conservative bias — unless Breitbart gets top billing.
By insisting on strong standards for sourcing, Facebook aims to improve transparency and thwart the spread of “fake news” about inflammatory topics. It will be interesting to see how the social network’s editors handle the thousands of stories by “reputable” news organizations that rely on unnamed sources and insiders.
It’s also heartening to see Facebook rely on human editors in an age of automation that includes news reporting, although they shouldn’t get too comfortable. Curating jobs beg for automation, given how little value they add to news publishing.
The automation wave already has begun.
Five years ago, the Associated Press announced a plan to automate its reporting of quarterly earnings from public companies. That was a relief to any reporter who has experienced the soul-deadening chore of churning out boilerplate earnings stories that don’t get any website hits. I’ve been there.
It won’t be long before artificial intelligence and natural-language processing technologies become advanced enough to replicate the decisions about what to curate. That’s fine, as long as those algorithms are open to public scrutiny to address claims of bias.