Good told the commission that He emphasized that Meta does not track antisemitic speech separately from overall hateful conduct, making it impossible to confirm any specific rise in anti-Jewish hatred using the company’s data. Good also said Meta has updated its hate speech policy to remove posts that use “Zionist” as a proxy for Jews or Israelis in a dehumanizing or violent manner, demonstrating continued attention to antisemitism.

Meta’s internal metric for hateful conduct violations has remained steady at 0.02% from 2022 through the period following the January 2025 policy changes, which he said indicates no measurable increase in hate content on the company’s platforms.

Meta’s January 2025 overhaul included a reduction in censorship, ending its third-party fact-checking program in key markets, and broadening the range of speech allowed on Facebook and Instagram. The company scaled back automated detection and removal of hate speech, reserving proactive enforcement mainly for severe categories such as terrorism and child sexual exploitation. Instead, Meta said it would rely more heavily on user reporting to address “lawful but awful” content, including hateful speech that does not meet the threshold of illegality under local law. To replace fact-checking, Meta introduced a “community notes” system modeled on the platform X, allowing users to append context or rebuttals rather than having posts removed. The Meta Oversight Board later criticized the rollout as “hasty” and lacking sufficient regard for human rights impacts, including the risk of increased hate targeting protected groups.

Meta representatives told the inquiry it was “entirely unrealistic” to attribute changes in moderator behavior solely to the policy changes, pushing back against claims that the moderation shift directly unleashed antisemitic content. Good described the company’s approach as a “reactive process” for handling hate speech, stating that Meta will respond to reports of hateful content but will not systematically scan for “lawful but awful” speech. Internal enforcement data presented to the Royal Commission showed a 79% drop in the volume of hateful conduct cases Meta took action on after the policy change. Good argued this reflected enforcement decisions rather than an underlying rise in hate content and maintained that the shift from proactive to reactive enforcement does not equate to tolerating more hate. He said the rollback was intended to reduce “over-enforcement” and protect freedom of expression, not to loosen standards on antisemitic or other hateful speech.

However, independent research and civil rights organizations have reported contrasting findings. The Anti-Defamation League (ADL) documented that antisemitic harassment of Jewish U.S. lawmakers on Facebook nearly quintupled after the moderation changes, rising from an average of 6.5 antisemitic comments per day to almost 30. The ADL’s analysis covered more than 337,000 comments on the Facebook pages of 30 Jewish members of Congress between February 4 and April 7, 2025. The organization attributed the surge largely to Meta’s moderation rollback and concluded that the shift to user-driven reporting and reduced proactive enforcement “bears responsibility” for enabling antisemitic and hateful activity. ADL researchers also found that Meta’s trust-and-safety changes have produced a “dramatic increase of antisemitism and extremism” on its platforms, warning that Instagram risks becoming a hub for hate. Civil rights groups and independent analysts reported that Jewish creators face increased antisemitic trolling and harassment, while content countering extremism sometimes faces automated takedowns under the new system.

The backlash from Jewish advocacy organizations and lawmakers was swift following Meta’s January 2025 announcement. A December 16, 2025, letter from U.S. Senators Ben Ray Luján, Jacky Rosen and Martin Heinrich to Meta CEO Mark Zuckerberg expressed concern that the policy changes were exposing Jewish communities to more antisemitic abuse and demanded detailed data on hate speech enforcement. The ADL issued a formal statement expressing deep concern about Meta’s trust and safety changes and warned that outside a few categories such as terrorism and child sexual exploitation, the company would no longer proactively enforce its policies. The ADL’s recent “report card” found that Meta responded to “exactly zero” cases of antisemitic content when reported by average users in its testing, highlighting a significant gap between policy promises and enforcement outcomes. The Meta Oversight Board criticized the company for hastily implementing the January 2025 changes with insufficient consideration of their impact on vulnerable groups, noting that updated guidelines allowed broad negative statements about protected groups.

The Royal Commission was shown enforcement data indicating a 79% drop in Meta’s actions on hateful conduct cases, which commissioners suggested was plausibly linked to the announced policy shift, despite Meta’s attempt to decouple the two. Critics argue that Meta’s reliance on a single global prevalence metric masks localized spikes in antisemitic content and fails to capture harms experienced by targeted communities. The tension between Meta’s official stance and independent findings centers on a fundamental disagreement over measurement and responsibility: Meta denies causation based on its internal data and framing of “lawful but awful” speech, while researchers, Jewish organizations and oversight bodies cite empirical studies and lived experience to argue that the 2025 changes have materially increased antisemitic content and harassment.