ع
The draft law builds on Australia’s world-first ban on social media for those under 16 that took effect late last year, according to reports from The Guardian and The New York Times. Under the proposal, platforms must proactively notify users over 16 via pop-up messages to choose between personalized algorithmic recommendations or a chronological feed of posts only from accounts they follow. Users could switch their preference repeatedly, and the measure applies to services including Instagram, TikTok, Facebook and YouTube as well as search engines and AI chatbots. The bill also directs platforms to prevent the spread of material promoting eating disorders, misogynistic content from the so-called manosphere, illegal pornography and videos that glorify dangerous stunts.
Communications Minister Anika Wells framed the changes as shifting responsibility from individual victims to the companies operating the platforms, a Nine.com.au report stated. Wells told parliament that the reforms would ensure a safe online environment by requiring annual risk assessments and giving the eSafety commissioner stronger powers to issue content removal orders. In an earlier briefing covered by NPR, Prime Minister Anthony Albanese said the legislation would give people genuine and enduring choice about what appears in their feeds after years in which technology firms had operated without sufficient oversight. The minister added that platforms must keep the opt-out tool easy to access and not degrade the experience for those who disable algorithmic recommendations.
A professor of internet studies at Curtin University in Perth noted in a New York Times interview that social media originally relied on reverse chronological feeds from followed accounts before recommendation engines took over to maximize engagement. The Australian approach goes further than the European Union’s Digital Services Act, which requires platforms to offer a non-profiling option but does not mandate repeated proactive prompts to users, according to analysis published by The Next Web. Government figures show the penalties for failing to comply could exceed A$100 million, roughly $72 million, as a deterrent for companies such as Meta, ByteDance and Google.
Australian Institute of Criminology figures show that 27.5 percent of respondents saw violent fringe and radical content online in 2024, more than double the 12 percent recorded in 2022.[[1]](https://www.aic.gov.au/publications/sr/sr62) The same assessment found algorithms played an increasing role in both first-time and repeated exposure to such material. eSafety Commissioner research found that 63 percent of children encountered content associated with harm in the past year, most often on platforms including YouTube and TikTok. An assessment by economist Andrew Leigh linked increased social media use to a 50 percent rise in young people reporting mental disorders between 2007 and 2022.
The proposal has drawn mixed reactions from opposition politicians, a Guardian report indicated. Liberal Party leader Angus Taylor expressed skepticism, arguing the measure risked amounting to government censorship of online content. Shadow communications minister Sarah Henderson separately criticized the psychological manipulation enabled by recommender systems, particularly for younger users, highlighting internal tensions within the Coalition. Wells has said expert advice would guide final details on whether the default setting requires an opt-in or opt-out for algorithmic feeds before full implementation after a planned 12-month transition period.
Personal accounts of users trapped in harmful loops by recommendation engines have fueled support for the reforms, according to coverage that appeared across outlets including the BBC. Advocacy groups including Teach Us Consent had pressed for the changes after documenting how engagement-driven algorithms amplified damaging material, a Daily Aus summary showed. The government plans to highlight the legislation at the United Nations General Assembly later this month as part of a broader push for international standards on platform accountability. Industry experts have warned that implementation details will determine effectiveness, with some platforms already offering hidden settings to view followed content only.
