How Meta's Content Rules Are Under Scrutiny in Australia's Antisemitism Investigation

Benjamin Good, Meta's global director of core policy, faced detailed questioning on 6 July 2026 at Australia's Royal Commission on Antisemitism and Social Cohesion. The commission's counsel, Richard Lancaster, pressed Good on a critical question: did Meta's decision in January 2025 to scale back content moderation open the door to more antisemitic material on Facebook, Instagram, and Threads?

The royal commission was formally established on 9 January 2026 in response to a terrorist attack in Bondi on 14 December 2025. Since its first public hearing in Sydney on 4 May 2026, the commission has been investigating how social media platforms govern content and what role they play in the rise of antisemitism in Australia. The 6 July session marked the first time Meta's policy architecture faced direct scrutiny at the inquiry.

What Meta Changed, and Why

In January 2025, Meta announced a major shift in how it moderates content. The company would reduce proactive moderation — the practice of scanning for and removing harmful material before users report it. It eliminated partnerships with third-party fact-checkers and narrowed enforcement to two categories: illegal content and what Meta called "very serious violations."

Mark Zuckerberg framed this as a necessary trade-off. "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," he said, according to The Guardian.

At the inquiry, Lancaster reported that the commission had received complaints from people saying the changes had allowed more antisemitic content to circulate. He also pointed to a concrete concern: witnesses who had testified at the royal commission were subsequently targeted with harassment on Facebook — suggesting the policy shift had real consequences.

What Meta Actually Allows

The hearing revealed excerpts from Meta's internal policy documents, which spell out exactly what the company permits and what it removes. The documents show that Meta allows certain kinds of offensive statements. Posts saying "gay people are sinners" are permitted. Broad generalisations like "immigrants are criminals" or "white people are all Nazis" are allowed. The company also permits statements such as "Black people are more violent than whites."

The line Meta draws is between general statements and specific claims about criminal behavior. A post saying "black people are all drug dealers" crosses that line because it attributes a specific crime to a group. The company's internal FAQ puts it plainly: "It is not Meta's role to police offensiveness."

This distinction — between offensive speech and specific incitement to harm — is central to how Meta operates after its 2025 policy change. It is also what the royal commission is most closely examining.

Meta's Defense: The Cost of Over-Policing

Good did not dispute the accuracy of the internal documents. Instead, he argued that removing the policy was necessary because aggressive content moderation creates its own damage: legitimate posts get taken down by mistake.

He pointed to Jewish communities as an example. During crises, members of the Jewish community had seen their own posts removed while speaking out against atrocities. Posts criticizing Hamas — a listed terrorist organisation — had been inadvertently deleted. Good called these outcomes "very bad" and said they illustrated why Meta had to pull back.

Good also told the inquiry that Meta's measurement of hateful content violations has remained steady at 0.02% of all posts since 2022. He presented this as evidence that the policy change had not led to a spike in hateful material.

The broader context here is important. The 0.02% figure may not tell the full story. This kind of metric measures violations that Meta's detection systems actually catch, not all hateful content that exists. If Meta reduces its proactive scanning, it will detect fewer violations — not because there is less hateful content, but because it is looking for it less actively. The number might stay the same while real-world harm increases.

What Happens Next

The 6 July session occurred during a technically dense stretch of the inquiry. On 2 July, the commission heard expert testimony that far-right actors are using coded language and memes specifically designed to evade automated detection systems, according to ABC News. A day later, the commission examined Australia's regulatory framework for social media. Witnesses told the inquiry that a mandated duty-of-care regime — a legal framework that would hold platforms accountable the way broadcasters are — remains years away from implementation, per ABC News.

That regulatory gap is the core problem. Good's testimony, whatever its merits, ultimately comes down to this: Meta regulates itself using rules it created. The political moment that established the royal commission — and its terms of reference — suggests Australian policymakers are losing patience with that arrangement. Whether the final report will produce enforceable recommendations is the question the remaining hearings will help answer.