Delays in online safety reform and gaps in the law came under renewed scrutiny at the Royal Commission on Antisemitism and Social Cohesion on Friday as Federal officials were pressed over the slow pace of action against harmful online content.
The commission heard from Sarah Vandenbroek, First Assistant Secretary in the federal Department of Infrastructure, and Theo Hourmouzis of Anthropic Australia Pty Ltd during a day focused on digital platforms, regulation and artificial intelligence.
Vandenbroek was questioned at length about the government’s response to the 2024 Rickard review of the Online Safety Act. The commission heard that a proposed digital duty of care for social media platforms is still at least 18 months away, despite growing concern about online hate and abuse.
She was also pressed on the absence of a definition of online hate in the legislation. The commission heard that legal advice has not yet been sought on what such a definition should include and that the department had also not sought legal advice on whether major platforms should be required to maintain a local presence in Australia.
In one of the day’s sharper exchanges, counsel assisting Richard Lancaster took Vandenbroek to X Corp’s responses to a departmental survey on harmful online content. The commission heard that X selected “none of the above” when asked what harmful material platforms should take steps to prevent for children, and again when asked the same question in relation to adults. Vandenbroek described that response as “disappointing”.
She also accepted that platforms can have a financial incentive to leave inflammatory material online because it drives engagement and advertising revenue.
The commission also heard there are no settled plans to broaden the eSafety Commissioner’s role or significantly increase funding, despite repeated evidence this week about the scale of online abuse and antisemitic material.
Later, the inquiry heard from Anthropic Australia representative Theo Hourmouzis, as the commission widened its focus to artificial intelligence and how AI systems respond to hateful prompts and harmful content.
Friday’s evidence sharpened a central theme of the current Sydney hearing block: online hate is spreading faster than the legal and regulatory systems meant to contain it.