Australia Pushes Meta Censorship Clampdown

Australia’s antisemitism royal commission is now pressuring Meta to censor more speech online, exposing a growing global clash between free expression and government-backed “hate speech” controls.

Story Snapshot

Meta told Australia’s antisemitism inquiry it shifted from proactive to reactive moderation, and removals of “hateful conduct” posts fell 79 percent.
Commissioners highlighted Holocaust denial and antisemitic posts that stayed up for years, raising doubts about Meta’s new enforcement model.
Meta now allows harsh phrases about immigrants, white people, and gay people, saying it will not “police offensiveness.”
The fight reflects a broader push by Western regulators to demand more removals, while platforms and many conservatives warn of mounting censorship.

Royal Commission Pressures Meta Over Drop in Removals

Australia’s Royal Commission on Antisemitism and Social Cohesion called Meta leaders to explain why hate-related removals collapsed after a major policy change in January 2025. Meta’s own figures showed action on 5.8 million Facebook posts in late 2024, dropping to just 1.2 million by mid‑2025, a roughly 79 percent fall in “hateful conduct” enforcement. Royal Commissioner Virginia Bell pressed Meta policy director Benjamin Good to link the drop to the new approach, but he refused to confirm direct causation and said he did not want to speculate.

Good told the inquiry Meta had moved from “proactive” moderation to a mostly “reactive” system that relies on users reporting problem posts. In the past, Meta used artificial intelligence tools to find and remove some harmful content before most people saw it, especially in areas like terrorism and serious crime. Now, for lower‑severity issues such as hateful conduct, Meta waits for user flags, arguing this cuts the risk of “over‑enforcement” and mistaken removals of lawful speech. Good still described proactive moderation as the “gold standard,” but admitted the reactive model can never fully match it.

Meta Admits It Will Catch ‘Less Bad Stuff’ to Protect Free Speech

At the heart of this dispute sits a clear trade‑off that Meta’s own chief executive, Mark Zuckerberg, has openly acknowledged. In a 2025 announcement about the new rules, he told users the changes “mean we’re going to catch less bad stuff, but we’ll also reduce the number of innocent people’s posts and accounts that we accidentally take down.” Meta’s enforcement update later boasted about “more speech and fewer mistakes,” claiming a roughly 50 percent cut in moderation errors in the United States between late 2024 and early 2025. For many free‑speech advocates, that sounds like a needed course correction after years of Big Tech overreach.

To reach that goal, Meta raised the confidence thresholds its automated systems must meet before they can act on lower‑severity policies, shifting the burden onto user reports. The company insists harmful content that attracts views will still be flagged by users and removed, while fringe posts with few viewers are less of a systemic risk. Critics at the commission argue this places too much faith in reporting and ignores vulnerable victims who may be too afraid or discouraged to complain. Yet Meta counters that aggressive algorithms can silence people who are speaking out against hate, especially within minority communities.

Holocaust Denial and Antisemitic Posts Expose Enforcement Gaps

Commission lawyers used real‑world examples to challenge Meta’s claim that the new model remains effective at stopping antisemitic abuse. One Holocaust denial post stayed on Instagram for more than three years, from 2020 to 2023, despite six user reports. Meta admitted only two of those reports were ever reviewed by human moderators, a stunning failure given the subject matter and the company’s pledge to crack down on Holocaust denial. During the hearings, four antisemitic Facebook posts appeared in real time; they were initially judged non‑violations and removed only after further review.

These cases gave the commission ammunition to argue that relying on user reports and looser rules leaves dangerous content online far too long. Good responded that it was “always difficult” to know why specific posts were missed and promised to learn from the incidents. He urged the inquiry to focus on metrics like total views of violating content instead of isolated mistakes, saying Meta targets high‑impact posts first. Still, with global antisemitism a rising concern, the image of Holocaust denial lingering on a major platform for years undermines public confidence in Meta’s assurances.

‘Not Meta’s Role’ to Police Offensive Opinions

Beyond specific failures, the commission also dug into Meta’s updated rulebook and found a broader change in how the company defines hate. Internal documents show Meta now allows phrases such as “immigrants are scum,” “white people are all Nazis,” and “gay people are sinners” on its platforms. Good explained that such comments may be offensive to many, but they do not always qualify as harmful conduct under Meta’s standards and that “it is not Meta’s role to police offensiveness.” Instead, the focus is on direct threats, harassment, and content clearly linked to real‑world harm.

Meta also told the inquiry it has no standalone policy aimed only at antisemitism; it relies on a mix of existing hate, violence, and misinformation rules to tackle Jew‑hatred. Good argued this approach is needed because antisemitism is “coded” and crosses into other topics, from geopolitics to religion, making a single rule too narrow. That explanation did little to satisfy campaigners who want explicit protections and faster removals. Groups like the Center for Countering Digital Hate claim Meta’s changes could gut up to 97 percent of its hate speech enforcement, and they are urging regulators to force a tougher line.

Global Fight Over Speech, Tech Power, and Government Control

This Australian battle fits a wider global pattern where governments and advocacy groups demand stricter online policing while platforms argue for freer speech and fewer mistakes. In the United States, lawmakers on the left push “online hate speech” laws, while many conservatives warn that vague bans on hateful or harmful words quickly turn into censorship of political, religious, and cultural views. Research on content moderation shows these policies often sweep up the speech of ordinary people, including members of minority groups discussing their own experiences, even as truly dangerous content still slips through.

Meta now controls the news feeds of about 3.5 billion users across Facebook, Instagram, and Threads, giving it huge power over what people see and say. At the same time, royal commissions like Australia’s have only advisory power and cannot force Meta to reverse its policy shift. That leaves a tense stalemate: regulators highlight failures, victims describe serious abuse, and Big Tech insists that more censorship will only create new injustices. For liberty‑minded Americans watching from afar, the lesson is clear. Any call for broader “hate speech” controls abroad will eventually echo here at home, threatening our First Amendment traditions unless we stay alert, demand transparency, and defend the line between true threats and unpopular opinions.