Platforms can’t measure antisemitism

The world’s largest social media platforms faced sustained scrutiny at the Royal Commission on Antisemitism and Social Cohesion this week, with representatives from Meta, TikTok, YouTube and livestreaming platform Kick confronted with evidence of significant gaps in moderating and removing antisemitic content.

Across two days of hearings in Sydney on July 6 and 7, Commissioner Virginia Bell heard a consistent picture. Platforms publicly commit to combating hate speech but in practice leave users and community organisations to flag harmful material, with inconsistent enforcement and limited data on the scale of the problem. None of the major platforms could tell the commission from their own data whether antisemitism on their services was increasing or decreasing.

Earlier in the hearing block, the commission heard that antisemitic content across social media had increased approximately fivefold since November 2023, with calls to violence spiking after the Bondi Beach attack.

Meta global director of core policy Benjamin Good, appearing voluntarily via video link from the United States, told the commission the company does not maintain a standalone antisemitism policy and does not track hateful conduct by the group it targets.

Action on hateful conduct on Facebook fell about 79 per cent after changes announced by chief executive Mark Zuckerberg in January 2025 shifted enforcement from proactive automated detection to a reactive model reliant on user reports, with a comparable 73 per cent fall on Instagram. Good accepted Meta already has the AI tools to detect antisemitic content at scale, but said the company wanted to deploy them “responsibly”.

Commissioner Bell tested Good’s claim that user reporting “tends to work well” by noting that four antisemitic posts reported to Meta in association with the commission had initially been assessed as non-violating before being removed on review. Antisemitic abuse directed at the commission’s own witnesses was similarly found not to breach Meta’s policies, and removed only after the commission wrote to Meta’s Australian lawyers.

The commission also heard a Holocaust denial post on Instagram remained on the platform for approximately three and a half years despite being reported six times before Meta’s oversight board intervened in January 2024.

Facebook Australia director of public policy Mia Garlick said Facebook Australia is not responsible for content moderation or removal, and could not provide data on antisemitism trends among Australian users.

On day seven, TikTok global head of policy, trust and safety, Zachary Hecht, was taken through the case of an antisemitic comment posted on May 7, which he acknowledged was “clearly a violation” of TikTok’s community guidelines. The content was correctly blocked by TikTok’s automated system but reinstated by a human moderator after the user appealed. When the Dor Foundation subsequently reported the comment, reviewers again assessed it as not in breach. It was only removed on July 2 after escalation to senior trust and safety staff.

TikTok global head of partnerships, elections and market integrity, Valiant Richey, confirmed the platform has no content moderators based in Australia despite employing around 760 staff and 16 contractors in Sydney and Melbourne. TikTok’s community partner channel achieves removal rates above 90 per cent, but evidence from the Online Hate Prevention Institute showed only 22 per cent of antisemitic content was removed through standard user reporting.

YouTube government affairs and public policy manager for Australia and New Zealand, Rachel Lord, was questioned about a video in which four men discussed the Bondi Chanukah attack, described survivor Arsen Ostrovsky as a “front man” and repeatedly labelled the incident a “false flag”. The video had been reviewed at senior levels and found not to breach YouTube’s community guidelines. Asked what would happen if it were reported again, Lord conceded “absolutely nothing would happen”. Before CyberWell became a priority flagger with YouTube, only 10.9 per cent of its reports resulted in removal.

Melbourne-founded livestreaming platform Kick took action on fewer than one in 100 reports of hate, doxxing and terrorism this year, the commission heard, with just 1774 of 179,914 reports resulting in action between January 1 and May 27. General counsel of parent company EasyGo Group, Tiat Oon Ooi, was unable to say with confidence whether a stream declaring “Jews are evil rats and subhumans” would breach the platform’s hate speech rules.

CyberWell founder and chief executive Tal-Or Cohen Montemayor told the commission only 50 per cent of antisemitic posts her organisation flagged were removed. She urged platforms and regulators to build dedicated, categorised data sets on antisemitism, warning that without transparency as “the baseline”, enforcement would keep failing at scale.