JTA — In January, a Turkish Instagram user posted part of an interview with Ye, the US artist formerly known as Kanye West, making statements in praise of Hitler and denying the Holocaust as well — and swiftly had the post removed from the social media platform.
But that wasn’t a success story for Meta’s efforts to keep hate speech off Instagram, the company’s Oversight Board has ruled. That’s because the Ye interview was accompanied by a reaction video of a reporter condemning his comments and sharing a family link to Holocaust victims.
Meta’s hate-speech detection system failed to grasp that the user was criticizing Ye, not endorsing him. So the company removed the post, accusing the user of violating its hate speech policies.
Now, the Oversight Board, an independent body tasked with reviewing Meta’s content moderation decisions, says the owner of Facebook and Instagram should improve its efforts to distinguish posts that promulgate hate from ones that aim to combat it. Too often, the board explained in a summary of its decision, human and automated moderators flag posts that are meant to educate against hate and antisemitism.
“Such mistakes can suppress speech meant to respond to hate speech, including Holocaust denial, or condemn statements of praise for dangerous individuals such as Hitler,” the summary said. “Protecting counter-speech is essential for advancing freedom of expression and a tool for combating harmful content such as misinformation and hate speech.”
Meta’s Oversight Board said Meta disregarded the context within which the user presented Ye’s comments and erroneously removed the post under policies barring Holocaust denial and praise of figures such as Hitler. The user appealed the removal to no avail, but when the Oversight Board brought the matter to Meta’s attention, the company acknowledged the error and restored the post.
The Oversight Board decided to examine the case despite the correction because it considers the underlying issue important, it said in a summary of its reasoning.
The board reiterated recommendations made in previous cases that Meta check how often its content moderators and algorithms incorrectly remove posts meant to educate about or counter hate speech. The accidental censoring of educational content has plagued the company for years, spiking, for example, when Meta banned Holocaust denial in 2020.
It is not the first time the Oversight Board, set up in 2020 amid allegations that Meta’s platforms help the spread of misinformation and extremism, has taken up a case involving antisemitism. Of the nearly 50 cases that have passed through its docket, at least four have touched upon Holocaust denial or Nazi-related content.
The most recent case came last month when the board announced it would examine Meta’s handling of a post distorting the Holocaust. In earlier cases, the board overturned Meta’s removal of a post featuring a quote incorrectly attributed to Nazi propagandist Joseph Goebbels and a post comparing the Russian military to Nazis.