Australia’s “My Feed, My Way” Bill: Algorithm Opt-Out, Platform Duties and Enforcement

Australia’s proposed “My Feed, My Way” initiative would give social-media users a meaningful choice over how their default feed is selected: accept algorithmic recommendations or switch to a feed based on people and groups they have chosen to follow. It is one feature of a proposed Digital Duty of Care, a wider reform designed to make online services responsible for managing foreseeable harms—not a blanket prohibition on algorithms.917

What users would be able to choose

Under the draft proposal, social-media platforms would have to notify new and existing users and offer a choice over their default feed. Users could opt in to a personalised feed that includes algorithmically recommended content, or opt out of those recommendations.715

Reporting on the proposal describes the opt-out feed as content from accounts, creators, individuals and groups the user has actively chosen to follow. In practical terms, that would reduce the role of recommendations based on a user’s demographics, interests or previous activity—including posts from accounts they do not follow.19

That distinction matters: the plan is framed as user control over feed ranking, rather than a government order to remove a defined class of speech.

The broader Digital Duty of Care

“My Feed, My Way” is only one part of the proposed framework. The Australian government has said the Digital Duty of Care would be legislated under the Online Safety Act 2021. Its stated purpose is to require online services to take reasonable steps to prevent foreseeable online harms.17

The proposed model would require regulated services to maintain effective systems and processes to provide, as far as reasonably practicable, a safe online environment. Government and legal-policy material describes obligations focused on preventing, monitoring and addressing illegal and harmful content, and on the safety of service features such as AI systems, algorithmic recommendation systems and bot accounts.

The practical shift is from a model in which users bear much of the burden of managing their own exposure to one in which platforms must proactively identify risks, mitigate them and sustain their safety measures. Reported policy goals include reducing exposure to harmful content and limiting the amplification of dangerous or divisive material by recommender systems.110

Why the government is targeting recommendation systems

The government has presented the reform as a follow-on to Australia’s under-16 social-media restrictions and as a way to give Australians more control online.715

Recommendation systems can surface material from outside a person’s chosen network and keep serving similar content based on predicted engagement. The proposal does not prohibit that model. Instead, it would require platforms to give users a route away from it. Whether this materially reduces compulsive or harmful use will depend on the final design: how prominent the choice is, whether it persists, and what content a non-personalised feed actually contains.

How it compares with the EU Digital Services Act

Australia’s approach is comparable to European platform regulation at a high level: it focuses on platform systems, foreseeable risks and mitigation rather than treating every problem as a question of removing an individual post. The Australian government previously described its proposed duty-of-care direction as aligned with approaches in the United Kingdom and European Union.

But it should not be described as an Australian copy of the EU’s Digital Services Act. The evidence available here does not establish a provision-by-provision equivalence. Australia’s proposal would operate through its own Online Safety Act framework, while its particularly visible user-facing element is the proposed choice to opt out of personalised recommendations.17

Enforcement and possible penalties

The Digital Duty of Care is proposed within Australia’s online-safety framework, where the eSafety Commissioner is the relevant regulator. Reports on the released draft say platforms would need to document harm-reduction measures and could face penalties of up to A$109.2 million for non-compliance.8

Because the measure is still a draft, the final enforcement architecture and penalty settings remain subject to consultation and parliamentary consideration. The clearest conclusion is that non-compliance is intended to carry substantial financial consequences; readers should avoid treating every reported detail as settled law until a bill is introduced and enacted.11

Separate reforms have strengthened enforcement of Australia’s under-16 social-media regime. Those measures would allow eSafety to compel companies to provide evidence of their compliance steps and increase the maximum penalty for breaches of that age-ban regime to A$99 million. Those powers and penalties should not automatically be assumed to be identical to the final Digital Duty of Care rules.

Free-speech and censorship questions

The algorithmic-feed choice itself is less directly censorial than a content ban: a user who opts out is choosing not to receive personalised recommendations, rather than the government ordering platforms to suppress a category of speech.

The broader duty of care raises harder questions. A legal requirement to prevent foreseeable harms could create pressure for platforms to remove or down-rank lawful but controversial material to reduce regulatory risk. The final impact will depend on how “harm” is defined, what guidance eSafety issues, what transparency and review mechanisms apply, and what protections exist for political, journalistic and minority speech. Those safeguards cannot be assessed conclusively from the material currently available.

Status and timeline

Draft Digital Duty of Care legislation was released on 8 September 2026 for targeted consultation, so “My Feed, My Way” remains a proposal rather than an enacted requirement.15

Earlier consultation informed the framework, including discussion of user control such as the ability to disable recommender systems. The provided material does not establish a firm date for the government to introduce a final bill to Parliament after this consultation. For now, the key point is straightforward: Australia is proposing a user-controlled alternative to personalised social feeds alongside a much broader obligation for online services to manage foreseeable harm.17