Australia Wants an Off Switch for the Algorithm: What the 'My Feed, My Way' Bill Would Actually Do

On October 14, 2026, legislation went before Australia's parliament that would do something no country has yet tried at this scale: force the world's biggest social media companies to hand users a real off switch for the algorithms that decide what they see. The proposal, branded "My Feed, My Way," is the centerpiece of the Albanese government's Digital Duty of Care bill, and it arrived amid a fierce parliamentary fight over how far the law should go.

What "My Feed, My Way" actually requires

The core requirement is simple to describe and hard to build. Platforms would have to notify new and existing users and offer them a choice over their default feed, according to a government statement reported by the Associated Press when the draft was unveiled in September 2026. Australians over the age of 16 would be able to opt in to a personalized, algorithmically ranked feed, or choose instead to see content only from friends and creators they actively follow, arranged without profiling-based recommendation.

The choice would have to be genuine and enduring, not a toggle buried three menus deep. That wording matters because the bill's drafters are clearly aware of how platforms handle user-choice features elsewhere. The European Union's Digital Services Act has required very large online platforms to offer a non-profiling feed option since August 2023, and regulators on both sides of the Atlantic have spent years debating whether those options are prominent enough to count. Australia's version would make the offer a statutory duty of care rather than a product-design preference, which is a meaningful legal escalation: a duty implies the platform is responsible if the choice turns out to be illusory.

The bill treats the algorithm like a default utility setting: the user should be able to decline it, and the platform should not be able to make declining it a chore.

Why Australia is going first

Canberra has spent the past year building a reputation as the world's most willing laboratory for platform regulation. In December 2025 the country enacted a world-first law barring children under 16 from holding social media accounts on major platforms, a policy that has since prompted similar moves in the United Kingdom and France. The Digital Duty of Care bill is the next step in that campaign, and it reaches further than the algorithm switch alone.

The draft laws would also require companies such as Meta and X to screen out harmful content for under-18s, including pornography, content promoting eating disorders, material promoting hostility toward women, crime, dangerous stunts, and content causing serious mental health distress, according to Sydney Morning Herald reporting. The underlying idea is to import the concept of a duty of care from product-safety law into the online world: platforms would be legally obligated to take reasonable steps to protect Australians from foreseeable harm, with enforcement in the hands of the eSafety commissioner and penalties of up to 109.2 million Australian dollars (about US$78.6 million) for failure to act. As of October 2026, only the eSafety commissioner could sue or fine companies under the draft; individual users would have no direct right to bring a case.

The parliamentary fight: private lawsuits and percentage fines

Passage is far from assured. The Coalition and One Nation have called the legislation a form of censorship, with opposition communications spokeswoman Sarah Henderson describing the proposed laws as dangerous and undemocratic and calling on Labor to scrap the draft and start over. With the Coalition unwilling to negotiate, the government will most likely need the Greens' support in the Senate.

The Greens, led on this issue by Senator Sarah Hanson-Young, want the bill to go further rather than softer. Their proposed amendments would create a private right of action, meaning individual Australians who believe they were harmed online could sue big tech companies directly for compensation instead of relying on the eSafety commissioner to act. The Greens also want stronger guardrails on addictive algorithms and fines tied to a percentage of technology companies' global revenue, arguing that the current 109.2 million dollar ceiling would not change behavior at firms that earn billions by keeping users on their platforms. Communications Minister Anika Wells has acknowledged that many Australians value algorithmic feeds, but she has argued the decision should rest with users rather than platforms; she was due to address the National Press Club about the draft laws on October 15, 2026. One concession is already on the table: the government is preparing to drop the ministerial discretion to designate new harms, a provision critics across the spectrum said handed the minister too much power.

What it means for the tech industry

For platform operators, an enforceable algorithm off switch would create a new kind of compliance surface. The feed is the product: ranking systems decide what users see, what creators get reach, and what advertisers buy. A prominent, persistent, legally required opt-out changes the economics of engagement, because a meaningful share of users given a real choice may take the chronological feed and leave behind the personalization data that powers ad targeting.

The industry already has a preview of this regime in Europe, where the DSA's non-profiling requirement has applied to very large online platforms since August 2023, and Meta's implementation lets people in the EU view Reels and Stories in chronological order and run non-personalized searches. But Australia's approach pairs the opt-out with a statutory duty of care and commissioner enforcement, a combination that raises the stakes beyond the European model. The tension with Washington is already visible: a White House economics spokesman, Kush Desai, has warned trading partners against imposing digital services taxes, fines, and other measures on American technology firms, which he characterized as extortion. Meanwhile Meta recently agreed to an $18 billion US settlement with some American states and territories to resolve claims the company harmed children, a reminder that legal risk around platform design is compounding on several fronts at once.

Analysts tracking the sector expect that if Australia's law passes in something close to its current form, regulators in other countries will study it as a template for user-control mandates, the same way the under-16 ban was copied abroad. That is the pattern Canberra is betting on: legislate first, and let the rest of the world decide whether to follow. As of October 15, 2026, the bill is proposed legislation and nothing in it is in force; the final shape is still being negotiated in the Senate, with the Greens' amendments on private lawsuits and revenue-based fines as the main variable to watch.