Australians could soon be given the right to switch off personalised social media algorithms and return to feeds made up primarily of posts from accounts they actually follow, under sweeping online safety reforms being prepared by the Albanese Government.
The proposed change will form part of Labor’s long-awaited Digital Duty of Care framework, which is designed to shift responsibility for online harm away from individual users and place a proactive legal obligation on technology companies to identify, prevent and reduce foreseeable risks created by their platforms. An exposure draft is expected to be released this week, with the government working towards legislation later this year.
Under the proposal reported by the ABC, social media platforms would be required to regularly ask users, through pop-up messages, whether they want to continue using an algorithmically curated feed or instead see content from the people and accounts they have chosen to follow.
The change would amount to one of the most direct attempts yet by an Australian government to regulate the recommendation systems at the heart of platforms such as Instagram, Facebook, TikTok, X and YouTube.
Algorithms determine which posts, videos and recommendations appear prominently in a user’s feed. Rather than simply displaying content chronologically, platforms analyse information including viewing behaviour, engagement patterns and other user data to predict what is most likely to keep a person watching, clicking or scrolling.
Those systems have increasingly become a political target because of concerns that content which provokes anger, fear, outrage or insecurity can attract greater engagement and therefore be repeatedly promoted.
Prime Minister Anthony Albanese has previously warned that algorithm-driven feeds can progressively expose users to more extreme material. In June, he said people could begin with relatively mainstream content and eventually find themselves being served extremist propaganda and calls for violence.
The government’s approach would not abolish personalised algorithms. Users who want algorithmically recommended content could keep it.
Instead, platforms would be required to provide a meaningful alternative, with users periodically reminded that they can switch away from personalised recommendations.
That distinction is already becoming one of the central political arguments surrounding the reform.
Campaigners including Teach Us Consent founder Chanel Contos have called for a stronger “opt-in” model, under which the algorithm would be switched off by default and users would have to actively choose to activate personalised recommendations.
Contos has argued that an opt-out system is an important first step but may not go far enough, particularly for teenagers beginning to use social media. Her Fix Our Feeds campaign says personalised algorithms should operate on the basis of informed choice rather than being automatically imposed when someone opens an account.
The Greens have also pushed for algorithms to become an opt-in choice.
Senator Sarah Hanson-Young introduced the Online Safety Amendment (Fix Our Feeds) Bill 2026 into the Senate in April. Parliamentary records show the bill seeks to amend the Online Safety Act to give social media users the ability to opt out of recommended algorithmic content, while the Greens have publicly argued that personalised feeds should be off by default unless users choose them.
The government’s plans, however, extend far beyond the algorithm switch.
The proposed Digital Duty of Care represents a fundamental change in how Australia regulates online safety.
At present, much of the system relies on users reporting abusive, illegal or harmful material after it appears. Under the new framework, companies would instead be expected to assess potential harms before they occur and build systems to reduce those risks.
The government’s May policy paper said online providers would be required to maintain effective systems and processes that, as far as reasonably practicable, provide Australians with a safe online environment.
Platforms would also have obligations to prevent, monitor and address illegal content and material considered harmful to young people, while ensuring that platform features including artificial intelligence, recommendation algorithms and bot accounts are designed and operated safely.
Importantly, the duty is being designed around systemic failures rather than making platforms automatically responsible every time a harmful post appears.
The government says companies would not necessarily breach the law because of a single instance of harmful content. But repeated failures, widespread harmful material or an inadequate response to known risks could provide evidence that a company had failed its duty of care.
According to the ABC’s report on the latest draft, the reforms would require companies to take steps to prevent adults being exposed to illegal material, including illegal pornography.
For people under 18, the framework is expected to identify additional categories of psychosocial harm, with body image and bullying among the areas being considered.
The eSafety Commissioner would sit at the centre of the enforcement regime.
The government’s earlier policy paper proposed giving eSafety powers ranging from information-gathering notices and formal warnings to remedial directions, enforceable undertakings, audits, infringement notices and applications to courts for civil penalties.
The latest proposal would also reportedly allow the eSafety Commissioner and registered independent researchers to use so-called “sock puppet” accounts, including accounts posing as children, to examine what social media algorithms actually recommend to young users.
Such testing could allow regulators to assess a platform as an ordinary child would experience it, rather than relying solely on information provided by the technology companies themselves.
The government has previously acknowledged the importance of independent researchers being able to study how platforms operate, including potential access to data from high-risk services.
The financial consequences for companies that systematically ignore their obligations could also be substantial.
The ABC reports companies could face fines exceeding $100 million for breaches. The government’s May discussion paper had already indicated that penalties for serious and systemic breaches should be comparable with the largest penalties available under Australian competition law.
The reforms come less than a year after Australia introduced its world-first social media minimum age regime.
Since 10 December 2025, major platforms including Facebook, Instagram, Snapchat, TikTok, X, Reddit and YouTube have been required to take reasonable steps to prevent Australians under 16 from creating or maintaining accounts.
eSafety reported that platforms had removed, deactivated or restricted access to about 4.7 million accounts assessed as belonging to Australians under 16 by mid-January. The existing age-restriction regime can expose companies to penalties of up to $54.6 million where they fail to take reasonable steps to comply.
In March, the government tightened the rules defining which platforms are caught by the minimum-age regime, placing particular emphasis on features such as account-based recommendation algorithms, infinite scrolling, likes and other feedback systems, and disappearing or time-limited content.
The Digital Duty of Care would go considerably further.
The government’s consultation framework proposes applying the duty not only to conventional social media but to a wide section of the digital economy, potentially including messaging services, online games, dating services, pornography platforms, generative AI services, hosting providers, internet service providers, search engines, app stores and operating systems.
Communications Minister Anika Wells has repeatedly argued that online companies should be required to design safety into their products rather than responding only after people are harmed.
Earlier this year, while discussing online risks to children, Wells said the Digital Duty of Care would put the responsibility on digital companies to proactively protect users and ensure some of the world’s most powerful technology businesses were held accountable.
The Coalition has yet to settle its response to the latest proposal.
Opposition Leader Angus Taylor said the Coalition wanted parents to have greater control over what their children see online but expressed concern that the legislation could become a mechanism for censorship.
“I am deeply sceptical,” Taylor said, while stressing the opposition had not yet seen the government’s proposal and would examine the details when they were released.
There have also been different views inside the Coalition.
Shadow Communications Minister Sarah Henderson has previously criticised what she described as the psychological manipulation of children through harmful algorithms, while Shadow Treasurer Tim Wilson has argued that algorithms often reflect users’ existing interests and has warned against government intervention that could restrict legitimate opinions.
The debate is therefore likely to focus not simply on whether algorithms can cause harm, but on who should control them and how far government regulation should extend.
Supporters argue users currently have little genuine understanding of how recommendation systems influence what appears on their screens and that platforms have powerful financial incentives to maximise engagement.
Critics are likely to demand strong protections for freedom of expression and safeguards against governments or regulators determining which lawful opinions people can see.
The government’s framework says those concerns will have to be taken into account. Rules created under the Digital Duty of Care are proposed to be subject to parliamentary scrutiny and would need to address compatibility with human rights, including freedom of expression.
For Australians using social media, however, the most immediately visible reform could be surprisingly simple: a message appearing on their screen asking whether they want the algorithm to keep deciding what they see.
If the laws proceed, Australians who answer “no” could increasingly find their social media experience looking more like the internet of a decade ago — posts from people they deliberately chose to follow, rather than an endless stream selected by an algorithm designed to predict what will keep them scrolling.
That apparently small choice sits at the centre of a much larger battle over the future of social media: whether technology companies should remain free to decide what fills Australians’ feeds, or whether the final decision should belong to the user.
Support our Journalism
No-nonsense journalism. No paywalls. Whether you’re in Australia, the UK, Canada, the USA, or India, you can support The Australia Today by taking a paid subscription via Patreon or donating via PayPal — and help keep honest, fearless journalism alive.
