Aviva's Submission on the Deepfake Digital Harm and Exploitation Bill
19th of June, 2026
About us
Aviva is a family and sexual violence charity established over 50 years ago. Our vision is a violence-free Aotearoa. Until then, we work to support people to overcome its effects and make families and communities safer. We work with people of all ages, genders and experiences of violence. It is only by breaking the cycles of abuse and trauma, which have plagued generations, that we will create safe futures for tamariki, in a world where all children can reach their potential. Since being established in 1973, we have worked with over 43,000 people. In the last five years alone, we have worked with more than 8,000 individuals/whānau, of whom 16% were tamariki. Since establishing our specialist services for survivors of sexual violence in 2014, we have supported over 4000 of such clients, and our referal numbers have increased each year, now accounting for around 40% of all whai ora.
Our position
This Bill is necessary, overdue, and supported across the political spectrum.
Aviva supports this Bill and we want it to succeed, which is why we are submitting amendments rather than opposition. Deepfake intimate imagery is image-based sexual abuse, and approximately 98% of deepfakes online are pornographic and target women.
The committee can substantially strengthen the Bill by adopting the changes described below. The amendments we propose below are the ones consistently called for by Cassandra Mudgway at Canterbury; Andrew Lensen at Victoria; Netsafe; the Privacy Foundation NZ; and the Equal Justice Project. They are:
- lower prosecution thresholds
- stronger platform obligations
- and supply-side regulation of the tools that make it too easy and inexpensive to create deepfakes
Lower prosecution thresholds
Under the existing Harmful Digital Communications Act offence, victims must prove the defendant intended to cause 'serious emotional distress' and that this actually resulted. That requirement is, in practice, a difficult hurdle for victim-survivors, and it shifts the focus of the inquiry to the victim's response rather than the inherent wrong of non-consensual sexualisation. The Bill should make clear that the new offence applies to the non-consensual creation, possession or distribution of synthesised intimate imagery without a serious-emotional-distress threshold being imported through the back door.
Outdated laws have left women in Aotearoa exposed and platforms unaccountable. The committee's report back should expressly characterise the conduct the Bill addresses as a form of gendered sexual violence. The framing matters because it determines how the offence is interpreted by courts and how victims are believed.
Further, the harms of non-consensual sexual deepfakes are significant regardless of whether the image is created from a real original or is entirely fabricated by a generative model (as recognised in a recent court case). The Bill's amendment to s 216N of the Crimes Act - defining the 'subject' as a person who is or 'appears to be' featured - is the right direction, and the committee should resist any narrowing amendment that would require the imagery to be derived from an existing photograph or to depict a perfectly identifiable real individual. The Classification Office's existing approach to AI-generated content - 'what matters is what the content shows, not how it was made' - should be the model.
Stronger platform obligations
Platform accountability is the part of this problem the Bill currently leaves on the table. Companies that deploy and host AI image-generation tools should be duty-bound to assess risk, implement effective guardrails, and prevent predictable misuse before it occurs - the voluntary Aotearoa New Zealand Code of Practice for Online Safety and Harms does not set standards for generative AI, does not require risk assessment, and does not impose meaningful consequences. The committee should recommend that the Bill be paired with statutory platform takedown duties and a published timeframe, so that victims do not have to obtain a court order before a platform will act.
Supplier regulation
The Bill addresses what someone does with a deepfake; it does not address the supply of the tools that produce them. Cassandra Mudgway has called for a targeted ban on consumer-facing nudify apps, following Australia, the United Kingdom and the European Union, together with mandatory safety guardrails in image-generation systems, stronger obligations on app stores and platforms distributing these tools, and transparency requirements around AI training data. Parliament should commit publicly to progressing that regulatory work in parallel with this Bill.
Recommendations
Our recommendations to the committee are as follows:
- Pass the Bill, and strengthen it with the survivor-centred amendments below.
- Recommend in parallel a targeted statutory ban on consumer-facing 'nudify' applications, following Australia, the United Kingdom and the European Union.
- Resource Netsafe, as the Approved Agency under the Harmful Digital Communications Act, to handle the increased complaint volumes that this Bill will and should produce, and ensure StopNCII and Take It Down style takedown pathways are promoted in any operational guidance accompanying the new offence.
- Ensure the new offence does not carry the existing Harmful Digital Communications Act 'serious emotional distress' threshold - the focus of the inquiry must be on the non-consensual sexualisation, not on the victim's measurable distress.