Product1 publisher3 min readPublished
Canberra's exposure draft takes a right Europe already granted and adds the part Europe left out: making platforms keep offering it. That obligation is also the rare online-safety rule a researcher can check in an afternoon.
The Product Desk · Product desk
Compiled by The Product DeskSomething wrong?How this is made
In Europe the largest platforms have had to offer at least one feed not built on profiling since the Digital Services Act of 2023 [5]. Take-up of the chronological feeds that resulted has been modest, and The Next Web attributes that to the DSA obliging nobody to put the option in front of a user [6]. Anyone who has shipped a settings screen to satisfy a rule knows the shape of it: the toggle exists and the box is ticked, while the count of people who ever found the toggle stays small.
What product teams tell themselves is that a user who wants different ranking will go looking for it. What users do is open the app and read what is on top. The Australian draft accepts the second description and moves the obligation from the existence of the choice to its delivery [1]. Plenty of jurisdictions have granted the escape route; The Next Web notes that almost none require the platform to keep asking [17].
That changes what a regulator has to prove. The under-16 ban is the local counter-example: Meta, TikTok and YouTube have been found not complying, and the government moved to double the fines and expand the regulator's powers in response [12], which is three of the biggest services in the country [16]. The Next Web's read is that Australia's age checks failed at their first test [13], and that a pop-up requirement is trivially auditable in a way age assurance is not, because an undercover researcher account can establish inside an afternoon whether the prompt appeared at the stated interval [14].
The 2x2 worth drawing has one axis for whether the non-profiling alternative exists in law, and one for whether anyone is obliged to surface it repeatedly. Europe sits in exists-but-unsurfaced [5][6]. The Australian draft moves to exists-and-surfaced while leaving the personalised default switched on, which is what an opt-out regime does: it relies on users acting [11]. The Greens want feeds to be opt-in, which flips the default rather than the prompt and sits in a different quadrant entirely [10].
Which quadrant you land in decides what compliance means. Under the draft, the artefact is the prompt. A platform can serve the dialog on schedule, watch almost everyone dismiss it, and remain inside the law, because the question the bill asks is whether the choice was offered rather than whether anyone took it. Under an opt-in rule the artefact is the feed itself [11].
The forcing function for a team with its own dormant setting: decide which number your obligation will be measured against before a regulator decides for you. If it is "the option is available", the work is a menu item and a screenshot. If it is "users chose it", the work is the default, and no amount of prompting substitutes. Australia has drafted the first instrument aimed squarely at the gap between those two, and it should be read as a delivery mandate rather than a new right [1][17].
Ranked by verification strength, evidence, and original report placement.
Australia published an exposure draft of the Digital Duty of Care bill this week that would require social media platforms to show users recurring prompts offering to turn the algorithmic feed off.
Reuters reported that the exposure draft carries penalties above A$100m for breaches.
The wider duty of care would require platforms to identify and mitigate risks on their services, prevent the sharing of harmful and illegal material, and take specific measures to protect minors from content relating to body image, bullying and other psycho-social harms.
The eSafety Commissioner and independent researchers would gain the power to monitor platforms using undercover accounts posing as children, a research method the companies have generally resisted.
The European Union's Digital Services Act, reached in 2023, requires the largest platforms to offer at least one recommender option not based on profiling, the same substantive right Australia is legislating.
Shadow communications minister Sarah Henderson has attacked the "insidious psychological manipulation" of children by recommender systems.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
One account, wire figure at second hand
Everything here reaches us through The Next Web. The politicians are quoted by name, and the duty's three limbs read as taken from the consultation text, but the draft itself is neither linked nor quoted, so the prompt's required interval and wording — the parts that decide whether this obligation bites — are absent. The single hard number, penalties above A$100m, is credited to Reuters, which means our coverage carries it without having seen the source document.
Consultation draft, nothing switched on
This is a consultation draft out for comment before it becomes a bill, and the chamber it enters has an opposition arguing with itself and a crossbench pushing for the stricter opt-in version, so nothing has taken effect yet. The only real-world signals point the other way. Europe's equivalent right exists on the major services and is described as lightly used, and the same Australian regulator has already found Meta, TikTok and YouTube short of the under-16 rules now in force.
Headline outruns the draft's status
The framing states that Australia will make platforms prompt users, and the body then concedes that this is a consultation document with an uncertain path through parliament. That is a modest stretch rather than a serious one, because the piece itself supplies the correction, spends its energy on enforcement mechanics instead of the announcement, and volunteers that Australia's previous flagship rule is being ignored by three of the services it targets.
Politicians on the record, platforms absent
Every quote comes from an opposition frontbench pulling in two directions on the same bill, with the Greens bidding higher from the crossbench, which is positioning as much as analysis. The government's own incentive is visible in the design: after age assurance failed publicly, a requirement that can be verified by an undercover account in an afternoon is a rule it can be seen to enforce. The companies carrying the A$100m exposure are not quoted or cited anywhere in this account, and the outlet's own stake is small -- a policy explainer with a newsletter sign-up at the foot.
Enough to follow, not enough to bank
The mechanism is described consistently and the political map is drawn from named speakers, so the shape of the draft can be trusted. Depth is thinner: one outlet, a second-hand penalty figure, an unquantified claim about European uptake, and no operative text for the prompt requirement. That is sufficient to track the story and insufficient for anyone sizing compliance work or exposure off it.
security
Meta's $17.1B settlement turns teen safety into an audited product spec8 publishers
science
Meta's $18bn settlement writes teen defaults into Instagram, with an auditor holding the pass mark4 publishers
product
France's under-15 ban failed on the age check, not the age limit1 publisher
product
Meta's settlement writes the teen spec: two hours, a midnight blackout, prompts every 15 minutes4 publishers
Publishers with included, body-backed reporting in this cluster.
1 article · September 7, 2026