On 15 October 2026, Communications Minister Anika Wells introduced Labor's digital duty of care bill. It would require social media platforms to protect users under 18 from content that promotes crime or eating disorders, from pornography and misogynistic content, and from posts that cause distress. The government wants it passed before the end of 2026, with commencement expected in late 2027, according to AAP reporting carried by the Canberra Times.
The strongest case for the bill
The argument for a duty of care is serious. Australia's existing framework, the Online Safety Act 2021, is largely reactive: it works through complaints and takedown notices after harm has occurred. A systemic duty shifts attention to how platforms design recommendation feeds and default settings, which is where much of the risk to teenagers arises. Labor describes the bill as baseline standards for large platforms, not a way to police individual users or individual posts. If that holds, it is a defensible design, and closer to product-safety regulation than to speech regulation.
What improved between the draft and the bill
The September draft would have let the minister define "serious harm" beyond the listed categories. The Nightly reported that critics, including Coalition communications spokeswoman Sarah Henderson, called that unfettered authority and a route to political censorship. Wells dropped the power; in her words, new harms can be added by Parliament through legislative amendment.
This is the most important free-speech safeguard in the package. A ministerial power to expand the list of restricted content, exercised by whoever holds the portfolio, would have turned a child-safety statute into a standing instrument of executive discretion. Requiring Parliament to act is slower, which is the point. It forces public debate, committee scrutiny and a record of reasons each time the scope of restricted speech grows.
The government also says it has included a carve-out for professional news material, and that users over 16 would get "My Feed, My Way" tools to opt out of algorithmic feeds. Choice over one's own feed is a proportionate, user-directed measure and far preferable to blanket removal duties.
Where the risk remains
The harm categories that remain are not equally precise. "Crime" and "pornography" have established legal meanings. "Misogynistic content" and "posts that cause distress" do not. A platform facing penalties of up to about $100 million, as Cyber Daily reported, has a strong incentive to over-remove anything near those lines. Other reporting puts the maximum slightly higher, so the exact figure should be confirmed against the bill text. The practical effect would fall on legitimate material: teenagers seeking recovery support for eating disorders, news coverage of crime, and sex education, all of which can be swept up by a cautious classifier.
A duty of care that depends on compliance judgments about "distress" invites exactly the defensive moderation that harms the users it aims to protect. The likely outcome is more age-gating, more automated filtering, and more identity checks, with real privacy costs for adults. Lawmakers should require platforms to show that their systems are proportionate and to publish how they handle false positives, not only how much harmful content they remove.
The private right of action question
Wells has said she will consider Greens proposals to let individuals sue platforms for harm. Greens senator Sarah Hanson-Young argues that without it the bill gives platforms little incentive to comply. The enforcement concern is legitimate: the eSafety Commissioner, Julie Inman-Grant, is expected to see her workload rise significantly, and a single regulator cannot supervise every platform.
But litigation is a blunt tool when the standard is "distress." Courts would have to decide, case by case, what counts as a harmful feed design, and platforms would respond by removing borderline speech to avoid liability. If Parliament adds private enforcement, it should limit it to demonstrable and serious harm, require a regulator finding first, and exclude claims based on lawful speech. A better answer to the capacity problem is to resource and structure the regulator properly and to give it transparent, appealable decisions.
Building on a contested base
The bill follows the world-first under-16 social media minimum age, set out in the Online Safety Amendment (Social Media Minimum Age) Act 2024, which amended the 2021 Act. Stacking a duty of care on an age-based ban means platforms will be assessed on how they handle minors under two overlapping regimes. Regulators should publish combined guidance, so compliance does not collapse into the cheapest option: verify everyone, restrict everything.
The parliamentary inquiry due to report in late November is the moment to fix this. Three changes would make the bill stronger without weakening its purpose. First, define the harm categories in the statute with objective thresholds, especially for misogyny and distress. Second, require independent evaluation of over-removal and a formal review after commencement. Third, keep the harm list in primary legislation, as the government has now committed to doing.
The government deserves credit for listening on ministerial discretion. The next test is whether it applies the same discipline to the vague standards, the penalty scale and any private right of action before the bill passes.