Australia is preparing to challenge one of social media’s most influential design choices: the personalized recommendation engine that decides what billions of people see.
Under a proposal known as “My Feed, My Way,” major platforms could be required to let Australian users choose between personalized algorithmic recommendations and a feed limited largely to accounts they have deliberately followed. The measure is being considered alongside Australia’s proposed Digital Duty of Care, a broader online safety framework intended to make technology companies anticipate and reduce foreseeable harms.
The idea sounds simple: add a social media algorithm opt-out. In practice, it raises complicated questions about ranking, discovery, advertising, content moderation and the reach of national technology laws. It has also drawn criticism from the U.S. government, which argues that some overseas platform regulations may burden American companies or indirectly restrict protected speech.
As of September 2026, the proposal remains subject to consultation, legislative drafting and parliamentary consideration. Its central direction is clear, but crucial details—including covered services, technical standards, transition periods and final penalties—are not settled.
What Is Australia’s “My Feed, My Way” Plan?
“My Feed, My Way” is Australia’s proposed answer to concerns that users have too little control over social media algorithms. Platforms routinely build individualized profiles from viewing time, likes, shares, searches, location signals and other behavior. Their systems then predict which posts will keep each person engaged.
Australia’s algorithm opt-out would require covered platforms to provide a meaningful alternative. Instead of being continually populated with recommended posts selected from across the platform, that alternative feed would focus on content from accounts, pages or channels the user has chosen to follow.
The proposal is not an outright ban on algorithmic feeds. Personalized social media feeds could remain available to people who want them. The policy’s central principle is choice: a user should be able to reject recommendation-based personalization without abandoning the service entirely.
For that choice to be meaningful, the non-personalized option would likely need to be easy to find, clearly explained and persistent. A platform that repeatedly resets users to its recommended feed, hides the setting behind several menus or uses confusing prompts could undermine the purpose of the Australia algorithm opt-out.
How the Digital Duty of Care Fits In
The proposed Australia Digital Duty of Care is broader than feed selection. It would move online safety regulation toward a preventative, systems-based model. Rather than responding only after harmful content reaches users, covered services would be expected to assess foreseeable risks and take reasonable steps to address them when designing products.
Potential areas of concern include features that amplify abusive material, self-harm content, eating-disorder content, violent material or other social media harmful content. Regulators could examine not only individual posts but also the product decisions that affect their distribution, including autoplay, notifications, recommendation engines and engagement-based ranking.
The Australian government’s stated goal is to place greater responsibility on platforms that create and operate these systems. The model is comparable to duties already found in fields where businesses must identify predictable risks before harm occurs.
Official background on the proposed framework has been published through the Australian government’s Digital Duty of Care consultation. However, the final legal relationship between that duty and “My Feed, My Way” will depend on legislation approved by Parliament.
What Would Actually Change When Recommendations Are Off?
Turning off social media recommendations would not necessarily produce a completely unfiltered or purely chronological stream. Every usable feed requires some rules for arranging content. The major difference would be the source of eligible posts and the signals used to rank them.
Discovery would become more intentional
Personalized feeds regularly insert videos, posts and creators that a user has never followed. With recommendations disabled, most of those suggested posts would disappear from the primary feed. Users would discover new accounts through deliberate actions such as searching, visiting an Explore page, following links or viewing content shared by existing connections.
This could reduce passive discovery, including the rapid exposure that helps emerging creators go viral. At the same time, it could give users a clearer understanding of why each post appears.
Ranking could still exist
A feed limited to followed accounts could still use an algorithm. For example, the platform might rank recent posts from close friends above older posts from brands. It could also group unread updates or prioritize live broadcasts.
The key legal question is whether the feed may use limited ranking for organization or must avoid behavioral personalization entirely. A strict version might require reverse chronological order. A more flexible version could allow ranking based on recency or user-selected preferences while prohibiting recommendations from accounts the user did not choose.
Suggested posts would be reduced or separated
Platforms may need to remove suggested posts from the following feed or place them in a separate, clearly labeled discovery area. Mixing frequent recommendations into an alleged opt-out feed would make the setting largely cosmetic.
The same issue applies to “people you may know,” trending topics and automatically inserted videos. Regulators will need to decide which features count as social media recommendations and whether optional discovery modules are permitted.
Content from followed accounts would matter more
Posts from selected accounts would no longer compete as heavily with content chosen from the entire platform. That could improve visibility for friends, publishers and creators whom the user intentionally followed. It could also reveal how little followed content appears in today’s algorithmic feeds.
Not every post would necessarily be shown. Spam controls, duplicate suppression, legal restrictions and user safety settings could still affect distribution.
Ads and moderation would probably remain
An Australia algorithm opt-out would not automatically create an ad-free experience. Advertising could still appear, although lawmakers must clarify whether those ads may be behaviorally targeted.
Content moderation would also continue. Platforms could remove illegal material, enforce their terms, hide blocked accounts and apply age-appropriate safeguards. Choosing a non-personalized feed would not give users a right to receive content that violates the law or a service’s published rules.
Why the United States Has Criticized the Proposal
The U.S. government’s objections sit within a wider dispute over foreign regulation of American technology companies. U.S. officials have warned that rules adopted by allies can become discriminatory trade barriers, impose substantial compliance costs or pressure platforms to restrict lawful American speech.
Three concerns are especially important in the U.S.-Australia social media dispute.
- Protected speech: U.S. officials are wary of government requirements that may incentivize platforms to suppress, demote or demonetize lawful viewpoints. American constitutional protections are generally broader than the speech protections available in many other countries.
- Content moderation pressure: A broad duty to prevent foreseeable harm could encourage companies to over-remove borderline material. Platforms facing major penalties may choose the safest regulatory option rather than risk a finding that their systems amplified harmful content.
- Impact on U.S. companies: The largest affected platforms are predominantly based in the United States. Washington may view prescriptive product rules, large fines or Australia-specific engineering obligations as disproportionately affecting American businesses.
The U.S. criticism does not necessarily mean that offering a feed choice is itself censorship. A user-controlled following feed can be framed as a product setting rather than a speech restriction. The sharper conflict arises when social media algorithm regulation is combined with broad duties concerning harmful—but potentially lawful—content.
Australia’s Goals and America’s Objections Are Not the Same Claim
Australia says the proposal is about social media safety, accountability and social media user control. Its argument is that engagement-driven systems can repeatedly amplify sensational or harmful material, particularly when users cannot easily escape personalized recommendations.
The U.S. objection focuses on legal spillover. Officials question whether vaguely defined harm standards could influence which speakers receive distribution and whether regulations aimed at global platforms place excessive burdens on U.S. firms.
Both positions can exist simultaneously. Recommendation systems can create genuine safety risks, while poorly drafted obligations can produce over-moderation. The eventual Australia social media law will be judged largely on whether it separates feed control from viewpoint control and provides precise, reviewable compliance standards.
How Platforms Might Respond
If the plan becomes law, platforms could add a prominent choice during onboarding and within feed settings. Users might select “Recommended” or “Following,” with an explanation of how each mode works.
Companies would also need to document whether the selection persists across sessions and devices. Regulators may expect evidence that the opt-out works as described, including audits of how many suggested posts appear in the following feed.
Other likely changes include:
- Separate tabs for personalized and following-only feeds.
- Clear labels identifying recommended content.
- Controls for chronological, recent or user-prioritized ranking.
- Explanations of the data used to generate recommendations.
- Reporting on recommendation risks and mitigation measures.
- Special defaults or protections for younger users.
Some platforms already offer following or chronological views. The regulatory difference is that those options could become mandatory, standardized and subject to enforcement rather than being removable product features.
Algorithm Transparency Will Be a Key Test
A button alone will not deliver meaningful social media algorithm control. Users must know what changes when they press it. Does the platform stop profiling their behavior, or does it merely stop using that profile in one feed? Are reels, notifications and search results still personalized? Can recommended content return after an app update?
Effective algorithm transparency may therefore require plain-language disclosures, accessible settings and independent oversight. It does not necessarily require platforms to publish proprietary source code. Regulators could instead demand information about input signals, recommendation objectives, testing methods and measurable outcomes.
Potential Fines and Enforcement
The Digital Duty of Care is intended to be enforceable rather than voluntary. Australia has already shown a willingness to use substantial civil penalties in online safety regulation, including maximum corporate penalties around A$49.5 million under related platform rules.
However, it would be premature to assume that every breach of “My Feed, My Way” will automatically carry that exact fine. The final bill must specify the penalty ceiling, whether penalties apply per violation, which regulator will investigate complaints and whether warnings or remediation orders must come first. A turnover-based penalty is another possible model, but it should not be treated as settled until Parliament approves the legislation.
Enforcement could involve user complaints, regulator-led investigations, transparency reports, compliance notices and court-imposed civil penalties. The law may also distinguish between a technical failure and a systemic decision to deny users the required choice.
Implementation Timeline and Unresolved Questions
No universal switch date should be treated as final while consultation and parliamentary consideration continue. The likely process includes completion of policy consultation, publication of draft legislation, committee review, debate and votes in Parliament, followed by regulations or regulatory guidance.
Platforms would probably receive a transition period to redesign feeds and compliance systems. Larger services may be able to adapt existing “Following” tabs, while smaller platforms could seek exemptions or simplified obligations.
Major unresolved questions include:
- Which social networks, video services and recommendation-based apps will be covered?
- Will the opt-out apply to all adults, children only or every Australian account?
- Must the alternative feed be chronological?
- Can ads, reposts and trending content appear in a following-only feed?
- How long must a user’s selection remain active?
- Will ranking audits be conducted by the regulator or independent specialists?
- How will Australian requirements interact with U.S. speech and trade concerns?
What the Proposal Could Mean Beyond Australia
Australia is part of a wider movement toward Big Tech regulation focused on platform design rather than individual posts. If its approach proves workable, other governments could adopt similar requirements for non-personalized feeds and recommendation transparency.
Platforms may decide that maintaining one global control is easier than building a separate Australian product. Alternatively, they could restrict the feature geographically, creating different social media experiences in different countries.
Either outcome would make Australia social media algorithms an international policy test. The debate is no longer only about what content platforms permit. It is increasingly about who controls distribution: the platform, the government or the user.
Frequently Asked Questions
No. The proposal is designed to provide a choice between personalized algorithmic recommendations and a feed centered on accounts the user has selected. Personalized feeds could continue for users who choose them.
Would turning off recommendations remove every algorithm?
Not necessarily. A following-only feed may still use rules to organize posts by recency or relevance. The key distinction is whether it introduces recommended content from accounts the user did not follow or relies on behavioral profiling.
Why does the U.S. government object?
U.S. concerns include possible pressure to moderate lawful speech, broad definitions of online harm, compliance costs and the disproportionate effect on U.S.-based technology companies. These objections are broader than the feed-choice feature alone.
When will the Australia algorithm law take effect?
A final commencement date has not been established. The proposal remains subject to consultation, legislative drafting, parliamentary approval and an expected implementation period.
Will users still see ads in a non-personalized feed?
Probably, although the final rules must clarify whether advertising can remain behaviorally targeted. The opt-out concerns content recommendations and does not automatically create an ad-free service.
The Bottom Line
Australia’s plan could give users something social platforms have historically resisted: durable control over the logic of the main feed. If implemented effectively, “My Feed, My Way” would reduce unsolicited recommendations and make content from chosen accounts more prominent without banning personalization for users who prefer it.
The difficult work lies in defining a genuine opt-out, protecting lawful expression and preventing safety duties from becoming vague incentives for excessive moderation. Until legislation passes, claims about exact fines, technical requirements or launch dates should be treated cautiously. What is already clear is that control over social media algorithms has become a significant point of tension between Australia’s safety agenda and the United States’ concerns about speech, trade and the treatment of American technology companies.
