The Digital Duty of Care Bill 2026 names pornography, and material that promotes hostile attitudes towards women, as harms to children. It does not name either as a harm to adult women.
The Online Safety Amendment (Digital Duty of Care) Bill 2026 names pornography, and material that promotes hostile attitudes towards women, as harms to children. Both are fixed directly in the Act: no minister’s decision is required for either to take effect.
The same two harms to adult women are treated differently, however. Neither appears in the list of material every platform must address for the general population. Both exist only as a possibility a minister could recognise later, by adding them to the list through a legislative instrument — a power written into the Bill itself that the minister may or may not choose to exercise, on no fixed timetable, and without the scrutiny an amending Act would receive. AAWAA’s submission on the exposure draft asks Parliament to name these harms for women in the Act itself.
Two kinds of duty
A government can hold a duty in one of two ways. It can treat the question as settled: the harm is named, the obligation is fixed, and platforms are not waiting on a future decision to know what is required of them. Or it can treat the duty as one claim among many, to be weighed against other priorities as they arise, with no guarantee that the weighing will ever conclude in the claim’s favour. The first is a form of guardianship, the second, stewardship. We have previously examined this distinction and its significance for our elected representatives in relation to the surrogacy sector and we extend that analysis here.
Section 25D of the Bill treats children’s protection from pornography and misogynistic material as guardianship. The question does not need to be revisited or justified each time it arises as it has already been answered. Section 25C(2), which governs whether the same two harms will ever be recognised in relation to women, is stewardship: recognition depends on a minister forming a view, at an unspecified time, in competition with everything else that might claim their attention. A duty that can always be deferred to the next decision is not, in any practical sense, a duty that has been accepted.
Why recognition alone is not the strongest argument
It would be possible to make this case more simply: pornography harms women, so it should be named in the Bill. That argument is true, but it is incomplete on its own, because it invites an answer the Bill already permits: agreement in principle, with action left for later. That is exactly the stewardship position section 25C(2) allows a minister to take indefinitely. The stronger claim has two parts, not one: that this harm is real, and that it is the kind of harm that should never have been left open to future, discretionary resolution in the first place. AAWAA’s submission makes both.
What actually justifies a guardianship duty
The comparison to children raises an obvious question. Women are not children. We are not legally incapacitated, and nothing in this argument depends on treating us as though we were. So what grounds a guardianship-style duty here, if not incapacity?
One answer suggests itself quite readily: that the harm is structural, produced at scale, and not something any individual woman could bargain her way out of. But that answer does not hold up on its own. Many harms are structural and difficult for an individual to escape without the affected group being owed anything resembling guardianship. If structural difficulty were sufficient on its own, almost any area of regulatory policy could be redescribed as a guardianship duty, and the distinction between a settled duty and a discretionary one would collapse into no distinction at all.
The stronger answer is finer than that. A guardianship duty is justified where a harm is not merely distributed across a class, but constituted by the exploitation of what that class is. Pornography’s harm to women does not happen to land on us incidentally, the way a rise in interest rates, for example, lands harder on people with large mortgages than on those who own their homes outright. Instead, the harm operates through the very fact of being a woman: production depends on it, and the material’s function depends on it. The characteristic is the mechanism of the harm, not an incidental feature of who it affects; in other words, being a woman is not a circumstance the harm happens to fall on — we are the essential material the harm is made from.
This is also the actual basis for how the Bill already treats children. Children are not owed guardianship because they are vulnerable in some general sense; plenty of adults are structurally vulnerable too, without a comparable duty attaching as already noted. Rather, children are owed it because certain harms, such as grooming and child sexual exploitation material, are designed around and depend on the fact of being a child. The shared ground between the two protections is not ‘both groups are vulnerable’ — it is that in both cases, the harm could not exist in its current form without the class membership it exploits.
The same test applies to material that promotes contempt for, hatred of, or the subordination of, women as a class, even though its harm takes a different shape from pornography’s. This material does not necessarily depend on women’s bodies being used in its production, but it does depend just as completely on women, as a class, being available to be held in contempt. Its entire function is to normalise a way of regarding women; if we strip out the fact of sex as a class, then there is no material left with anything to say. The harm is not incidental to the characteristic here, either: it is built from it, in the same way pornography’s harm is built from the fact of being a woman, and it meets the same guardianship test for the same reason.
This is not paternalism
A duty framed this way invites an obvious objection: that Parliament is deciding what is good for women, overriding any individual woman who says pornography has not harmed her, or who works in the industry and rejects being described as exploited by it.
That objection assumes guardianship and paternalism are the same thing. They are not. Paternalism assumes a class lacks the capacity to judge its own circumstances, and uses that assumed incapacity as the reason to override individual judgement ‘for its own good.’ Nothing in this argument makes that assumption, and nothing in it depends on any individual woman’s view of her own experience being wrong or unreliable.
The guardianship claim is different in kind. It says that some structures are exploitative at the level of the structure itself, regardless of what any individual participant in that structure reports about how she experiences it. An individual account of a particular experience is not evidence about whether the structure that produced it is exploitative: it is evidence about one person’s experience within it. The state can decide it will not endorse an exploitative structure without making any claim at all about the individual women who take part in it, just as it can decide a market is exploitative without needing every participant in that market to agree. That is the basis on which pornography can be named as a harm to women as a class, without the claim resting on, or being undone by, what any individual woman says about her own encounter with it.
What this means for the Bill
Naming pornography and misogynistic material as harms to women in the Act is not a request to be regarded as sympathetically as children, or to be treated the same by way of comparison. Instead, it is a claim that this harm meets the same structural test the Bill already applies elsewhere, and that meeting that test calls for the same legal treatment: fixed in the Act and not left to a minister’s later discretion.
That is what AAWAA’s submission asks Parliament to do. The Bill has already decided that some harms are too closely bound up with the exploitation of a class to be left open for later resolution. Pornography and misogynistic material both meet that description and both should be named on the same terms.
Read the full submission, below.