To curb online bullying, Instagram has to spot it

If you were to rank all the ways humans can inflict harm on one another, ranked by severity, it might be a few pages before you got to “intentional inducement of FOMO.”

Purposefully giving someone else FOMO — fear of missing out — is not a crime, or even a misdemeanor. But it is a big problem on Instagram, where millions of teenagers go every day to check on their peers. And it is one of the subtle slights that Instagram is focused on classifying as part of its new anti-bullying efforts, which will use a combination of artificial intelligence and human reviewers to try to protect its youngest users from harassment and pain.

It is part of a larger attempt by Instagram and parent company Facebook to clean themselves up. Both have struggled to contain a flood of toxic behavior, extreme content and misinformation.

Instagram is particularly vulnerable because of its young user base. About 70% of American teenagers use the service, according to the Pew Research Center. And 42% of cyberbullying victims ages 12 to 20 reported being bullied on Instagram, according to a 2017 survey by British anti-bullying organization Ditch the Label.

Instagram has discussed how to fight bullying for years, but there is a renewed focus. Last year, the Menlo Park company announced an effort to use AI to label instances of bullying within photos. This year, it said it would begin testing features for improving teenagers’ mental health, including the ability to hide “like” counts on posts.

“There are a lot of teens using Instagram, so we actually see new behaviors and words all the time, and we need to work quickly to understand if these new trends are harmful,” said Bettina Fairman, Instagram’s director of community operations.

These efforts are still unproven, and, like any Facebook-related promises, they’re best taken...