19 June | Lucia Moses | Digiday
Last week, The New York Times announced it would expand the use of automation to open up more articles to reader comments. Using a system called Moderator developed with Jigsaw, the tech incubator from Google’s parent company Alphabet, the Times hopes to expand the number of stories open to comments from 10 percent today to 80 percent by the end of the year. But the tech wouldn’t work without the foundation the Times community desk laid.
That’s because Moderator has been trained by being fed 16 million Times comments that have been hand-moderated over the past 10 years. It uses those judgments to build an algorithm that will prioritize comments for moderation.
The editor of the community desk is Bassey Etim, who leads a team of 14 part- and full-time moderators who review about 12,000 comments a day. It’s a complicated process, so the Times can only open up just so many articles to comments and only for a limited time.
Moderators aren’t just screening comments for curses and threats; they’re seeking to create a place where readers can discuss stories, exchange ideas at a high level and see all sides of a debate reflected. Etim calls it “content curation at a very high level.”
The job of the desk has gotten more complicated in the Trump era. There are just more comments, and in this heated political environment, it’s become harder to characterize what’s inflammatory.
People have gotten more creative about name-calling those they disagree with, whether it’s “liberal snowflakes” or “followers of the Great Orange,” leaving the Times to figure out how to reflect the culture without letting too much meanness get through.
A lot of how the Times handles these depends on context. It’s OK to say snowflake if someone is discussing liberals generally, but not to attack a specific commenter.
“Whereas things like ‘t-Rump’ and ‘Orange Menace’ are never allowed, just like we do not allow ‘DemoRats’ or ‘President O-Bozo,’” Bassey said.