Home › Cybercrime Law › Module 4 · Content offences

Module 4 · Content offences: child sexual abuse material and intimate image abuse

What this module covers and why it matters. This module covers the content offences. These are crimes defined not by an attack on a computer but by the nature of the material a person deals with online. It is the second of the offence modules and it sits in the subject's arc from offences to jurisdiction to investigations to cooperation. Three things are worth holding from the start. First, most content is not harmonised at all. States disagree about adult indecency and obscenity and no treaty binds them to a common line. Second, one content offence is harmonised. Child sexual abuse material is criminalised across jurisdictions and the Budapest Convention art 9 supplies the shared template. Third, a newer content wrong now sits beside it. Intimate image abuse covers the non-consensual sharing of sexual images including material created by technology such as deepfakes. The doctrinal load is heavy because these offences turn on precise definitional gates. The debate load matters too because the essay questions ask whether AI-generated material challenges the foundations of criminality and where the law should draw the create versus share line. This module is written soberly and at the level of doctrine and policy only.

How to use this guide. The Attack plan box gives you the order of analysis for a content-offence problem. The RULE cards state the operative provisions. The chips mark the definitional gates and the defence twist. The Debate section carries the research-essay load and each strand closes with critical-analysis questions. Read once to learn the map and then revise from the headings and boxes.

Instruments at a glance

Attack plan

The lack of harmonisation on adult content

What this section covers and why it matters. Before the harmonised offence it helps to see what is not harmonised. Adult indecency and obscenity are governed by national standards that differ widely and no treaty binds states to a common line. This section explains why that is so and why it sets the stage for everything that follows. The reason matters because material lawful in one state may be unlawful in another and the internet carries it across every border.

Start with the traditional position. The possession of obscene material was historically not an offence although production and distribution were (Clough ch 10). Most jurisdictions retain offences dealing with obscene or offensive material but the content of those offences is set by national standards rather than a shared rule (Clough ch 10). Prosecutions have been brought for offensive images of various kinds and for extreme pornographic material, hate speech and suicide-related or terrorism-related material, but the categories and thresholds vary from state to state (Clough ch 10).

The Budapest Convention leaves this to national law. It harmonises one content offence only. Where "pornographic material" is concerned art 9 provides that what counts as pornographic is to be determined according to national standards (Budapest Explanatory Report). The First Additional Protocol adds racist and xenophobic content offences but some states including the United States decline it on free speech grounds. So the treaty framework reflects the wider truth that states cannot agree a global standard for indecency.

Consolidation. Adult content is not harmonised. States set their own standards for indecency and obscenity. This is the backdrop against which child abuse material stands out as the one content offence the world has been able to agree.

The one harmonised content offence and its rationale

What this section covers and why. This section explains why child sexual abuse material is the single harmonised content offence and what justifies criminalising even its possession. It sits here because the rationale shapes how far the offence reaches, in particular to material that involves no real child. The reason matters because the AI-generated-material debate later turns on whether these rationales still hold when no child is harmed in production.

Child sexual abuse material is the one content wrong states have agreed to criminalise. The Budapest Convention identifies the need to modernise national laws so they address the use of computer systems in this offending and art 9 supplies the shared template (Budapest Convention; Clough ch 10). The international trade is a shared problem because material unlawful in one jurisdiction may be lawful in another and the global reach of the internet argues for consistency at least for distribution offences (Clough ch 10).

Criminalising possession needs its own justification because possession of obscene material was traditionally not an offence. The Supreme Court of Canada in R v Sharpe summarised the rationales the prosecution advanced for prohibiting possession and accepted that prohibition is linked to reducing the sexual abuse of children in five ways (Clough ch 10). Read them as the policy scaffold that supports the whole offence.

These rationales matter because they decide reach. The Supreme Court of Canada accepted that they apply even where no real child was used in production, which is how the offence extends to virtual material (Clough ch 10). The United States took the opposite view and this divergence is the seam the AI debate runs along.

Consolidation. Child abuse material is criminalised widely and possession is criminalised for reasons that reach beyond the abuse of a particular child. Whether those reasons justify reaching material made without any child is the contested question the module returns to.

Defining the material: the definitional gates

What this section covers and why. The offences all turn on whether the thing in question is child abuse material. This section sets out the definitional gates that decide it. It sits at the centre of the module because in a problem question the classification does most of the work. Each gate is a point where a defence may live.

The Budapest Convention supplies the template definition. Article 9(2) defines child pornography to include pornographic material that visually depicts one of three things (Budapest art 9).

Limbs 2 and 3 are what reach virtual material. A party may reserve the right not to apply limbs 2 and 3, so a state may confine its offence to material involving a real child (Budapest art 9(4)). Australia does not take that narrow path. Its definition reaches representations and so covers virtual material (Clough ch 10). Now take the gates in turn.

The age gate. Under the Convention a minor is a person under 18 years of age (Budapest art 9(3)). A party may adopt a lower age limit in defining the material so long as it is not below 16 (Budapest Explanatory Report). In Australia a minor for these purposes is a person under 18 (Clough ch 10). This produces a deliberate anomaly. A young person may be old enough to consent to sexual activity yet recording that activity may still be unlawful because the recording age is set higher (Clough ch 10). The relevant time for age is the time the image was made rather than the time it was viewed (Clough ch 10).

The "sexually explicit conduct" gate. The Explanatory Report states that sexually explicit conduct is intended to cover, whether real or simulated, sexual intercourse, bestiality, masturbation, sadistic or masochistic abuse in a sexual context, and the lascivious exhibition of the genitals or pubic area of a minor (Budapest Explanatory Report). At the margin, images that merely emphasise a child's genitals, anal region or breasts are caught only where they meet an added standard such as being indecent or lascivious or having a sexual purpose as the dominant characteristic (Clough ch 10). The test is objective. It asks whether a reasonable viewer looking at the depiction objectively and in context would see its dominant characteristic as the sexual depiction of the child (R v Sharpe, in Clough ch 10).

The "material" and "depicts" gate. In the digital environment the thing often exists only as data, so the definition must be technologically neutral. The Australian federal provisions simply refer to "material", defined to include material in any form capable of constituting a communication, which reaches data as well as photographs (Criminal Code s 473.1; Clough ch 10). The Australian definition also reaches material that "depicts a person, or a representation of a person", and it has been held that this ordinary meaning is capable of applying to cartoons of children (Clough ch 10). Contrast the older statutory terms such as "photograph" or "visual depiction" which required amendment to reach data files (Clough ch 10). The lesson is that "depicts" and "material" are drafted broadly in Australia so the medium of depiction does not create a gap.

Consolidation. Three gates decide whether material is caught. The age gate fixes who is a child. The sexually-explicit-conduct gate fixes what is depicted and applies an objective reasonable-viewer test at the margin. The material-and-depicts gate keeps the definition technologically neutral so data and representations are caught. A defence often lives at one of these gates.

The Australian carriage-service offences

What this section covers and why. This section sets out the operative Australian offences and their defences. It sits after the definitions because the offences all presuppose that the material has been classified. The heavy penalties and the narrow defences are the practical core of a problem answer.

The Commonwealth offences are carriage-service offences. They rest on the telecommunications power and criminalise dealing with the material through a carriage service. There are two parallel streams. The child pornography material stream is in ss 474.19 and 474.20. The child abuse material stream is in ss 474.22 and 474.23 and mirrors it (Clough ch 10). The 2019 reforms consolidated the definitions of the two classes in s 473.1, so the current single concept governs both streams. State and territory offences also exist and do not require use of a carriage service (Clough ch 10).

Consolidation. The Australian offences criminalise the full range of dealings with the material through a carriage service and carry heavy penalties. The defences are narrow and cast as public benefit and law enforcement exceptions on which the defendant bears an evidential burden. The age-based consent filter is the main mechanism that keeps the offences off consensual conduct between minors.

Producing and the meaning of "depicts" applied

What this section covers and why. This short section shows how "produce" and "depicts" work once the material is defined, because the same broad drafting that closes the medium gap also widens the producing offence. It sits here because producing is one of the listed dealings and its breadth surprises students.

The concepts of producing and making cover the obvious cases of taking photographs or video. They also reach making reproductions of existing images, using imaging software to create virtual material, and copying data from one storage medium to another (Clough ch 10). Downloading or copying data can itself amount to making the material, which is why the producing offences reach far beyond the original creator (Clough ch 10). The breadth follows directly from the technologically neutral definition of "material".

Consolidation. Because "material" and "depicts" are drawn broadly, "produce" is broad too. A person who copies or downloads may be producing. This is the doctrinal bridge to the AI debate, where the act of generating an image squarely engages the producing offence.

Intimate image abuse and the deepfake amendments

What this section covers and why. This section covers the newer content wrong, the non-consensual sharing of intimate images, and the 2024 deepfake amendments. It sits last because it postdates the textbook readings and the current statute governs. The reason it matters is that this is where the create versus share line is now being drawn in Australian law.

McGlynn and Rackley supply the framing. They conceptualise "image-based sexual abuse" as the non-consensual creation and/or distribution of private sexual images (McGlynn and Rackley 2017). They prefer that term to "revenge porn" for three reasons. The label "revenge porn" describes only a small subset, it fixes on the perpetrator's motive rather than the harm, and the word "porn" imports a false sense of choice and legitimacy (McGlynn and Rackley 2017). They situate the wrong on Liz Kelly's continuum of sexual violence and argue that its individual harms of physical and mental illness and the loss of dignity, privacy and sexual autonomy combine into a cultural harm that falls on society as well as the individual (McGlynn and Rackley 2017). Their concept expressly reaches images altered or "photoshopped" so that a person appears to have engaged in sexual activity, which they call sexualised photoshopping (McGlynn and Rackley 2017). The deepfake is the current form of that category.

Australia now criminalises this conduct through a carriage service. The offence covers transmitting private sexual material of a person who is or appears to be 18 or over without consent. The current statute postdates Clough (2015) and Urbas (2023) so the current provisions govern and the divergence is noted below.

The aggravated creation offence is the point of doctrinal interest. Section 474.17AA raises the penalty where the person both created or altered the material and transmitted it without consent. So Australian law now attaches a higher penalty to the creator who shares, but the base offence in s 474.17A still turns on transmission. Creation alone without transmission is not the gravamen of the carriage-service offence. This is exactly the line McGlynn contests in the deepfake debate.

Consolidation. Intimate image abuse is a distinct content wrong from child abuse material. It protects adults and turns on consent rather than age. The 2024 amendments drafted the offence to reach deepfakes and other technologically created material and added an aggravated offence for the creator who transmits. The base wrong is still non-consensual transmission, which frames the create versus share debate.

Common confusions

These are the errors that cost marks. Each states the mistake and then the correction.

Debate

This is the centre of gravity for the research essay. Each strand is a two-sided contest with the leading voices named. Each closes with questions to test a position rather than settle it. The prompts stay at the policy level.

Does AI-generated material challenge the foundations of the criminality? The rationales for criminalising child abuse material were built around the abuse of a real child. Some of the Sharpe rationales reach further and treat the material as harmful whether or not a real child was used, because it fosters and communicates the same degrading message and feeds the same market (R v Sharpe, in Clough ch 10). On that view virtual and AI-generated material can be criminalised without straining the foundations. The United States drew the opposite line. In Ashcroft v Free Speech Coalition the Supreme Court struck down the parts of the federal definition that reached material only appearing to be a minor, holding that the causal link between such material and actual abuse was too contingent and indirect to justify banning speech that harmed no real child (Clough ch 10). Congress responded with the PROTECT Act, narrowing the definition to material "indistinguishable" from a real minor, and the durability of that fix is contested (Clough ch 10). The morphed-image cases sharpen the line further. Where an identifiable real child's face is superimposed the courts have treated the material as closer to real abuse material because it harms an identifiable child, as in the reasoning in Bach and Anderson (Clough ch 10). AI generation stresses this framework because it can produce realistic material with no identifiable child at all, which pushes the question of whether the harm is the abuse of a child or the material itself.

• If no real child is involved, which of the Sharpe rationales still justifies the criminal law, and are they enough on their own?
• Should the law treat AI-generated material as equivalent to material involving a real child, or does the Ashcroft objection about contingent harm still bite?

Where should the law draw the create versus share line for deepfakes? McGlynn argues the criminal law should reach the creation of a sexual deepfake and not only its sharing, because the wrong and much of the harm occur at creation and a share-only offence leaves victims unprotected until distribution (McGlynn, in the image-based sexual abuse framing). Her wider work treats image-based sexual abuse as covering non-consensual creation as well as distribution and insists that terminology and thresholds fixed on distribution and on the perpetrator's motive leave gaps (McGlynn and Rackley 2017). Against a broad creation offence sits the difficulty of drawing the line. McGlynn and Rackley themselves note that people will disagree about how far criminal liability should extend to those who merely forward an image, and that governments have hesitated to criminalise "routine sharing" on social networks even though it is that routine sharing which produces much of the harm (McGlynn and Rackley 2017). Australia's answer is a middle position. Section 474.17A criminalises non-consensual transmission and s 474.17AA raises the penalty where the transmitter also created or altered the material, so creation aggravates a transmission offence rather than founding a standalone crime (Criminal Code ss 474.17A, 474.17AA). Whether that goes far enough for the person whose deepfake is created but not yet shared is the live question.

• Should creating a sexual deepfake be a complete offence in itself, or only an aggravating factor to non-consensual sharing?
• If secondary and hosting sharers cause much of the harm, where should liability stop before it captures ordinary social-media forwarding?

Can the world set a global standard for indecency? For child abuse material the answer has been yes, at least for a core, because art 9 supplies a shared template and most jurisdictions track it (Clough ch 10). For adult indecency the answer has been no. What counts as pornographic is left to national standards and states diverge sharply, so the same material may be lawful in one state and criminal in another (Budapest Explanatory Report; Clough ch 10). The reading guide asks whether the Convention should carry more content offences beyond art 9. The case for more harmonisation is that content crosses borders as freely as any other data and inconsistency lets material be hosted wherever the law is thinnest (Clough ch 10). The case against is the same objection that kept adult content out in the first place. States hold genuinely different views about indecency, harassment and harmful speech, and a treaty broad enough to bind them all may either fail to agree a meaningful standard or agree one so permissive of state control that it threatens legitimate expression. This strand connects to Module 1, where the free speech reservation to the First Additional Protocol shows the same fault line.

• Is child abuse material harmonisable because states share a value, or only because the harm is unusually clear-cut compared with other content?
• Would extending the Convention to further content offences raise the global standard, or expose the disagreement that keeps indecency national?

Discussion prompts

1. The Budapest Convention harmonises only one content offence. Consider whether it should be extended to further content offences beyond child abuse material and identify which if any could command genuine agreement.
2. Explain why adult indecency and obscenity have resisted a global standard and assess whether the difficulty is practical or a genuine disagreement about values.
3. Set out the definitional gates that decide whether material is child abuse material and explain which gate does the most work in a marginal case.
4. Evaluate whether the rationales for criminalising child abuse material still hold where the material is generated by AI and no real child is involved, drawing on the contrast between the Sharpe reasoning and the Ashcroft decision.
5. Assess the argument that the law should criminalise the creation of a sexual deepfake and not only its non-consensual sharing, and explain where you would draw the line of liability.
6. Consider how the image-based sexual abuse framing of McGlynn and Rackley would reshape the Australian intimate image offences and whether the 2024 deepfake amendments go far enough.