Behind the Lens: Social media star puts Instagram's anti-troll technology to the test

Last Updated Aug 15, 2017 2:45 AM EDT

"CBSN: On Assignment" airs Mondays at 10 p.m. ET/PT on CBS and on our streaming network, CBSN. To explore more on this topic click here.


With over 3.5 million followers on Instagram, Iskra Lawrence is a true social media star. While she has many fans, she is also no stranger to vulgar comments that appear on her Instagram photos. The 25-year-old model says she does not let those comments get to her, but she does worry about what effect such nasty comments might have on others who follow her.

"I'm more worried about those explicit words that young people are going to see," Lawrence said. "No one should type mean things."

Lawrence's celebrity quickly rose last year after a creative response to one "fat-shamer" went viral. It is hateful comments like the ones directed at her that Instagram wants to eliminate from the platform.

With more than 700 million active users worldwide, Instagram sees thousands of comments by the minute.  And in an effort to weed out the trolls, it has developed a feature that blocks the worst of the worst comments, using machine learning technology.

Lawrence accompanied CBS News to Instagram's New York offices to test out the new feature, after an early morning workout and photoshoot by the Hudson River. Using the photos taken at her morning workout session, the model joined Nicholas Thompson, editor-in-chief of Wired, and Julian Gutman, the Instagram product manager responsible for the algorithm used in the new feature, for a live demonstration.

iskra-lawrence.png

Model and social media star Iskra Lawrence (left) checks out her Instagram feed.

CBS News

"'You look fat and disgusting. I am going to throw up,'" Thompson typed into the comments on her page. "All right, so let's see if that showed up in your comments. It did make it! Julian, why did it get through?" Thompson asked.

Gutman says the comment passed through the filter because the feature is still "in the beginning" stages. Instagram is working on developing and refining the algorithm and will continue to do so over the next few months. Despite some shortcomings, the filter successfully deflected the majority of comments that used words like "slut," "whore" or "fat."

In a test with another ugly insult ("This is the internet! People say things like that," Thompson remarked) the algorithm did its job  and the comment never appeared in Lawrence's feed.

Lawrence believes that this effort helps make Instagram a really positive place. "I just breathe a sigh of relief… there's nothing really hurtful and offensive and you just want to make it a safe space."

"There is a way to disagree with someone but it doesn't have to be threatening," she said. "I've had people who didn't agree with what I've been doing, but they didn't need to say something explicit."

But she also feels that sometimes being able to interact with people who post negative comments on her Instagram is empowering. "I take back the power from the troll," Lawrence says.

Total censorship, she believes, is not necessarily the answer. "There's one guy and he calls me Miss Piggy… And I really don't mind. I kind of find it funny," she says.

How does Instagram account for the trolls we might like to engage with? Gutman says, "There's obviously a range of expression and we're really focused on the worst of the worst."