Instagram failed to act on 93% of abusive comments aimed at female politicians, study says

Instagram failed to remove 93% of abusive comments directed at female politicians out of 1,000 examples flagged to the app for most likely violating its rules, according to a report Wednesday from the nonprofit Center for Countering Digital Hate.

The center said it collected the abusive comments this summer from the Instagram accounts of 10 female officeholders — five Republicans and five Democrats. It said it reported the comments through the app and then checked back a week later to find the vast majority of them still online.

Many of the abusive replies Instagram left up were sexually degrading or included violent threats likely to be in violation of Instagram’s community guidelines, according to the center. Some comments included explicit language about rape, injury or death in connection with the female politicians, and Instagram left them up, the center said.

It’s not clear in the 93% of examples whether Instagram took action short of removal, for example by ranking abusive comments lower down in discussions so fewer people see them.

Meta, Instagram’s parent company, said it hadn’t seen the report before its release but would review the examples collected by the center. It said it would remove anything that violates the rules but added that not all offensive comments are violations.

The company said its systems are a work in progress.

“We provide tools so that anyone can control who can comment on their posts, automatically filter out offensive comments, phrases or emojis, and automatically hide comments from people who don’t follow them,” Cindy Southworth, Meta’s head of women’s safety, said in an emailed statement.

“We work with hundreds of safety partners around the world to continually improve our policies, tools, detection and enforcement, and we will review the CCDH report and take action on any content that violates our policies,” she said.

The report arrives at a potentially historic moment for women in politics, with Vice President Kamala Harris seeking to become the nation’s first female president. She’s expected to face former President Donald Trump in the November election.

Harris was among the 10 female officeholders whose Instagram accounts the center examined, and Instagram was no better at removing abusive language directed at her, the report found. The center said 92% of 105 abusive comments targeting Harris were still online a week after it reported them to Instagram. The abuse included sexist and racist comments, the center said.

The Harris campaign didn’t immediately respond to a request for comment on the findings. The offices of all the female politicians included in the report have been asked for comment.

The abusive comments and Instagram’s inconsistent enforcement illustrate the harassment women often face after they decide to run for office, said Imran Ahmed, CEO of the Center for Countering Digital Hate.

“They are seeking to send a message to women more generally: that you do not have a place in public life,” he said of the people leaving comments.

“Abuse of women can dissuade some from deciding that politics is the right career for them,” he said.

Ahmed said Instagram should enforce its existing policies, including bans on hate speech, violent speech and dehumanizing speech, and he said it should be more transparent with users about its reasoning when it rules on potential rules violations. 

He said the center focused the report on Instagram, rather than other social media apps, because it’s a platform that “people perceive as safe” and the center wanted to test whether that was true.

The abusive replies were still a minority of comments on Instagram, which has automated systems to catch some abuse and harassment. About 1 in 25 replies to female politicians on Instagram are likely to be toxic, according to the center’s analysis of a set of 421,361 comments.

Instagram says it catches 98% of the hate speech that it eventually takes action on before anyone reports it, thanks to its automated detection software.

But many of the abusive comments come from repeat offenders, according to the center. Of the 1,000 samples researchers examined, 221 comments were from accounts that had previously targeted the female politicians in the study, it said.