The Center for Countering Digital Hate called it “every parent’s nightmare.”

ByYi-Jin Yu via 
December 16, 2022, 4:59 AM

A new NGO report released Thursday claims TikTok’s algorithm targets “vulnerable teens” and recommends “harmful” content to them, sometimes as rapidly as every 27 to 39 seconds.

The report, from the Center for Countering Digital Hate, defines harmful content as content about topics like eating disorders, weight-shaming, self-harm, and sexual assault.

The center’s researchers conducted the study by signing up eight new accounts in the U.S., Canada, the U.K. and Australia, setting them to a minimum age of 13, and watching and liking videos that discussed body image and mental health. Half of the accounts had usernames with “loseweight” in them and were deemed “vulnerable” accounts by the researchers, while the other half were deemed “standard” accounts without a “loseweight” username.

“For each account, we recorded the first 30 minutes of algorithmically recommended content on each account’s ‘For You’ feed, watching and liking any videos about body image, mental health or eating disorders. The resulting recordings were analyzed to examine the frequency of recommendations for body image, mental health, selfharm and eating disorder content.”

According to the researchers’ findings, once the social media app’s algorithm kicked in, content about suicide surfaced on the “standard” accounts’ feed as early as 2.6 minutes in, and content about eating disorders popped up within 8 minutes. “Vulnerable” accounts, meanwhile, were recommended three times as many harmful videos and 12 times as many self-harm videos as the “standard” accounts.