Meta’s recent changes to its content moderation policies have not resulted in an increase in antisemitic content on its platforms, according to testimony given by the company’s policy director, Benjamin Good. Good spoke on Monday during a hearing before the Royal Commission on Antisemitism and Social Cohesion.

In January 2025, Meta implemented a significant overhaul of its approach to moderating hateful and harmful content. The company moved away from proactive enforcement measures, which involved automated detection and removal of flagged material, toward a system focused primarily on user reports and reactive enforcement. This shift was intended to reduce what Meta described as over-enforcement and to better safeguard freedom of expression across its platforms.

During the hearing, Good addressed concerns raised by Jewish advocacy groups who reported an uptick in antisemitic content on Meta’s services following the policy change. He described the company’s interactions with these groups as “overwhelmingly productive and constructive” and rejected the assertion that policy changes were targeted at easing restrictions on antisemitic speech. Instead, Good said the updates aimed to make incremental adjustments that would allow users to discuss political topics more freely, including issues debated in public forums such as Congress and the news media.

Good acknowledged hearing from members of Jewish communities who reported that, after October 7, some of their counter-speech efforts against antisemitism had been subject to content removals. He characterized these experiences as “extremely negative,” noting that the systems in place unintentionally impacted those speaking out against hateful content.

Data presented during the hearing showed a marked decrease in the volume of hate speech removals after Meta’s policy changes. Between October and December 2024, prior to the new approach, Meta removed 5.8 million items on Facebook classified as hate speech. By comparison, between July and September 2025, only 1.2 million such items were actioned, representing a 79% decline. When asked to explain this drop, Good declined to speculate, emphasizing that the switch to a reactive model was designed to mitigate over-enforcement rather than reduce efforts to combat hate speech.

He also agreed that some hateful content might remain unaddressed if not reported but said Meta is closely monitoring the effectiveness of its enforcement under the new system.

Separately, in April 2026, the Anti-Defamation League (ADL) criticized Meta for a perceived weakening of content moderation on Instagram after the platform adopted similar policy changes. According to the ADL’s research, Instagram failed to remove 93% of hateful and extremist content reported by its researchers, which the organization described as a “systemic failure” to protect users from harmful material.

Jonathan Greenblatt, CEO and national director of the ADL, said that Instagram is becoming a “hub for hate and antisemitism,” with the moderation rollback creating an environment where extremists can exploit platform features to spread hate, adversely affecting vulnerable communities.

This shift in moderation practices on Meta’s platforms follows a broader trend in social media, as platforms such as Elon Musk’s X have also reduced moderation rigor, allowing previously banned extremist accounts to return. The ADL has publicly criticized both Meta and Musk’s operations for their moderation approaches in recent years.