Why I Made a Submission on Australia’s Digital Duty of Care Bill

The Digital Duty of Care Bill

On 8 September 2026, the Australian Government released the Exposure Draft of the Online Safety Amendment (Digital Duty of Care) Bill 2026. The proposal would create a broad duty of care for online services used by Australians, with the stated aim of requiring providers to take greater responsibility for foreseeable online harms. Public feedback on the Exposure Draft closes at 12 noon on 22 September. I have now lodged a submission on the draft.

Those familiar with my work will understand why some of these concerns immediately stood out to me. I write from a traditional Christian worldview, and I am conscious that some of the convictions flowing from that worldview now sit uncomfortably with prevailing cultural assumptions.

Questions of biological sex, marriage, parental authority, the sanctity of life and religious belief can quickly become contested territory.

That does not place Christian views beyond criticism, nor should it. But it does make me particularly attentive to the point at which disagreement is reclassified as harm, hostility or a safety risk.

Across a number of articles I have examined questions of conscience, institutional power, regulatory expansion and free expression for precisely this reason. My concern is rarely limited to what a government says it intends to do today. I am equally interested in what a law makes possible tomorrow, once the machinery exists and the people operating it have changed.

My starting point is important. I do not oppose strong action against child sexual exploitation, grooming, genuine threats, terrorism, non-consensual intimate imagery or comparable harms. Children require serious protection online, and platforms should not be able to design dangerous systems and simply wash their hands of the consequences.

My concern begins at a different point: what happens when a framework built for objectively serious harms begins reaching contested attitudes, lawful public expression and broad assessments of what might constitute a risk?

When “harm” reaches contested ideas

One of the clearest examples is proposed section 25D. Among other things, it includes material or conduct that “encourages or promotes hostile attitudes towards women or gender equality” within the material regarded as harmful to children.

I have no interest in protecting threats, stalking, harassment, abuse or targeted intimidation of women and girls. Those are different questions. What stood out to me is that the legislation does not define “hostile attitudes”, nor does that particular category require targeting, threats or harassment. It reaches the promotion of attitudes.

That matters when public debate turns to questions such as biological sex, gender identity, women’s sporting categories and single-sex spaces, school curriculum, medical interventions involving minors, marriage, parental authority or religious teaching. People hold strongly conflicting positions on these issues. The fact that a view is offensive, unpopular or regarded by another person as hostile should not, by itself, transform lawful disagreement into a child-safety harm.

My submission therefore asks for greater precision. If Parliament intends to regulate serious harm, the boundary between serious harm and contested lawful expression should be clear enough for citizens, publishers, platforms and regulators to understand.

The protection that disappeared from view

Another issue stood out when I compared the Exposure Draft with the Government’s earlier May 2026 framework.

That earlier paper expressly recognised freedom of expression. It said services taking reasonable steps should have regard to legitimate uses and people’s right to freedom of expression so that those rights were not unduly impaired. It also described the proposed duty as being centred on risk assessments, mitigation strategies and effective systems and processes.

The Exposure Draft now requires the person responsible for an online service to “ensure, so far as is reasonably practicable, a safe online environment”. At the same time, freedom of expression is not expressly included among the considerations in the statutory test of what is reasonably practicable. That does not mean expression cannot be considered. The list is non-exhaustive. But privacy is expressly named, while lawful expression is left implicit.

For something as fundamental as public debate, journalism, religious expression and political communication, I do not think implication is the strongest safeguard.

My submission therefore recommends putting freedom of expression expressly into the statutory balancing exercise and making clear that restrictions on lawful public communication should be necessary, proportionate and no more restrictive than reasonably required.

What happens after the machinery exists?

The issue that runs most deeply through my submission is gradual expansion. Sections 25C and 25D allow additional categories of harmful material or conduct to be prescribed through legislative instruments where the Minister is satisfied that the material or conduct may cause the relevant harm.

There are safeguards around that process. Legislative instruments face parliamentary scrutiny and possible disallowance, and human-rights compatibility statements accompany disallowable instruments. I acknowledged those protections rather than pretending they do not exist, my concern is what happens cumulatively.

Once online services have risk assessments, content classifications, reporting obligations, remedial directions and substantial penalties incorporated into their compliance systems, a new category of harm does not enter an empty space. It enters an already functioning regulatory machine.

For regular readers, the reasoning here may already be familiar from pieces such as The Check We Never Built and A Familiar Thread. Both explored, from different angles, a question I keep returning to in my writing: not only what a power is intended to do when it is created, but what becomes possible once the machinery exists and is inherited by different governments, regulators and institutions.

That is why I have argued for greater parliamentary control where future categories would reach lawful public expression, along with published evidence, an assessment of the impact on freedom of expression and a shorter review period than the ordinary legislative-instrument cycle. This is not an allegation that the present Government has a hidden plan to suppress dissent, as it does not require one.

Governments change. Ministers change. Regulators change. Technology changes. Public understandings of words such as harm, safety and hostility also change. The question I keep returning to is therefore not simply, What does the Government say this power is for today? It is also, what is somebody able to do with the same power tomorrow?

Context still matters

Schedule 1 raises another issue that illustrates the importance of context. The draft broadens the treatment of material that has been reposted after already meeting a statutory removal threshold. Altered versions can include shortened or cropped material, screenshots, additions and changed commentary. I understand the purpose. A person should not necessarily be able to evade a lawful removal decision simply by cropping an image or slightly altering a post, but context can change meaning.

A screenshot used to continue abuse is not necessarily the same communication as that screenshot being reproduced by a journalist, researcher, commentator or citizen in order to criticise, report on or document what was originally said. The legislation therefore needs to preserve a distinction between evasion of a removal decision and legitimate discussion of the material that was removed.

The recent Baumgarten matter also informed my thinking about regulatory boundaries. eSafety has acknowledged that the complained-of material did not meet the statutory adult cyber-abuse threshold, but an informal complaint notification was nevertheless sent to X. The post was subsequently withheld in Australia. eSafety later changed its notification template to make clearer that such correspondence was not a legal notice.

I do not take that case to mean informal communication between regulators and platforms is inherently improper. eSafety itself has pointed to Justice Horan’s observation that regulatory functions may extend to requests without immediate legal consequences.

My concern is simpler: if an informal request can have a practical effect on lawful speech, the distinction between a voluntary request and a compulsory direction should be unmistakable.

Regular readers may recognise a related concern from The Accountability Gap, where I examined how informal referrals, professional regulation and institutional pressure can influence the information environment even where no direct legal order to remove speech exists.

Protection without treating disagreement as danger

At the heart of my submission is a distinction I think we need to preserve. A society can take genuine online harms seriously without pretending that disagreement, offence, controversy and serious harm are the same thing.

Children should be protected from exploitation. People should not be stalked, threatened or subjected to non-consensual sexual imagery. Platforms should take responsibility for genuinely dangerous systems they design and operate. But a free society must also retain room for people to disagree about questions that matter deeply.

My submission makes six recommendations directed at that boundary: expressly recognise freedom of expression, tighten unclear harm definitions, strengthen safeguards around future expansion, retain a systems-and-processes approach to platform responsibility, maintain clear boundaries around regulatory pressure, and make obligations proportionate to the services and risks actually involved.

I am not arguing that the current Government has announced an intention to suppress legitimate debate. I am asking whether the law contains enough safeguards for the governments, regulators and compliance cultures that will inherit these powers later.

A power should be judged not only by the purpose for which it is created, but by the capability it leaves behind.

If this proposal matters to you, I would encourage you to read the Exposure Draft for yourself and make your own submission before 12 noon on 22 September 2026. Whether you share my concerns or reach different conclusions, this is the point in the process when public feedback is being invited.

For those who would like to examine the argument and recommendations in full, I have made my submission available below.

Read my submission: Online Safety Amendment (Digital Duty of Care) Bill 2026

For me, this is also why freedom to question and test competing claims matters. Scripture does not ask us to fear examination; it recognises the wisdom of allowing a case to be heard, tested and challenged.

“The one who states his case first seems right, until the other comes and examines him.”— Proverbs 18:17

Thanks for reading.

God Bless.

Mark.