The recommendation algorithm has quietly become the thing people love and blame in the same breath, and Australia has decided to hand users a formal off switch. Draft legislation for a Digital Duty of Care, released by the Albanese Government today for targeted consultation, includes an Australian-first plan called “My Feed, My Way”: social media platforms would have to notify every user — new and existing — and offer them a choice over their default feed. Opt in to algorithmic recommendations, or opt out and see only the friends and creators you chose to follow. Under the draft, users over 16 get that choice, and platforms would be required to respect it.
Prime Minister Anthony Albanese framed it plainly: “This is not about giving government control, it is about giving people control.”
The AI chatbot clause nobody saw coming
The duty of care doesn’t stop at social media. The draft extends basic safety obligations to online games, apps and AI chatbots, which would be required to protect under-18s from design features with negative behavioural impacts — addictive loops, features that affect self-esteem — and from a defined list of harmful content, including material promoting eating disorders, pornography, and content glorifying crime or life-endangering stunts.
That puts conversational AI squarely inside a consumer-protection framework for the first time in Australian law. A chatbot’s engagement-maximising design choices would be documented, auditable, and enforceable. Communications Minister Anika Wells drew the comparison herself: “Just as we have basic safety standards for cars, toys or food, the duty of care applies basic standards for the online products we use every day.”
Enforcement sits with the eSafety Commissioner, and the draft gives the regulator new powers to issue removal notices for “nudify” apps and websites. Platforms would have to document the measures they take against identified risks of harm, and failure to comply could attract penalties up to $109.2 million. The Government wants feedback from platforms, industry bodies and civil society before legislation goes to Parliament later this year.
The harder experiment it follows
This is the second swing Australia has taken at platform design, and the first one is still finding its range. The world-first under-16 social media ban has struggled since taking effect — as we covered in August, eSafety’s own study found 81 per cent of teens were still using the platforms three months in, and half of children surveyed said their age was never checked. That enforcement gap is the backdrop to today’s release: age verification proved harder than legislating for it.
A feed toggle is a different kind of intervention. You cannot reliably verify a teenager’s age from a selfie, but you can verify whether a switch exists and whether a platform honours it. Of the two regulatory experiments, the opt-out looks like the one a regulator can actually police. It is also the one that treats adults as adults — the draft extends the choice to everyone over 16, not just to minors.
There is a catch worth naming. This is a consultation draft, so the mechanics that decide whether it works — how prominent the notification is, how many taps an opt-out takes, what the default feed degrades to without recommendations — have not been published. A choice buried five menus deep is not a choice. Whether platforms fight the obligation or quietly comply will depend on the penalty math, and $109.2 million is serious money even for them.
It also slots into a broader Australian pattern. The same government has already moved to constrain automated decision-making inside its own agencies, with a digital duty of care flagged as the follow-up — today’s release is that follow-up arriving, aimed at the private sector this time.
What it means for New Zealand
New Zealand has watched each stage of the Australian experiment closely — the trans-Tasman regulatory split matters to anyone working across both markets. There is no equivalent algorithm-choice proposal here, and no duty-of-care regime covering AI chatbots. If the Australian toggle survives consultation and works in practice, the obvious question for Wellington is whether a user’s right to a non-algorithmic feed should stop at the Tasman. Nobody here has proposed it yet.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
FAQ
Can Australians turn off the recommendation algorithm? Not yet. The choice is in draft legislation released 8 September 2026 for consultation. If passed, users over 16 could opt out of personalised recommendations in their default feed when opening the app.
What happens if platforms don’t comply? The draft would let the eSafety Commissioner enforce the duty of care, with penalties up to $109.2 million. The details could still change during consultation.
Does this cover AI chatbots? Yes — the draft duty of care extends to AI chatbots, games and apps, requiring them to protect under-18s from addictive design features and defined categories of harmful content.
