Will criminalising deepfake images and videos protect women and girls?
13th August 2025
In a disturbing development, Australian school students are increasingly distributing sexually explicit deepfakes of their peers.1 In January, the New South Wales education department referred a male student to police over allegations he used AI to create pornography depicting female students. A month later, Victoria Police launched an investigation into the circulation of deepfakes involving as many as 60 female students at Gladstone Park Secondary College in Melbourne’s northwest.2
According to eSafety Commissioner Julie Inman Grant, the number of deepfakes on the internet has increased by 550% since 2019.3 Of these, 99% are pornographic in nature and the victims are nearly always female. These startling statistics have prompted South Australian MP Connie Bonaros to describe imaged-based abuse as the ‘fastest growing threat to women and girls online today’.4
Until recently, there were no Australian laws directly addressing this disturbing trend.
The Online Safety Act 2021 (Cth) allows the eSafety Commissioner to impose civil penalties on anyone who posts or threatens to post an intimate image of another person without their consent.5 However, the Act only applies to real images, not deepfakes. The eSafety Commissioner nonetheless has the power to issue content removal notices to websites and social media platforms. But, as the eSafety Commissioner admits, they don’t always comply. Many of the offending publications are based outside Australia, have no point of contact, or hide their true hosting information.6
Then there’s the Privacy Act 1988 (Cth), which former Attorney-General Mark Dreyfus once described as unfit for the digital age.7 After widespread consultation, the Act was finally amended in November to allow the Office of the Australian Information Commissioner to impose civil penalties on individuals – not just agencies and organisations – who publicly disclose another person’s identity or private information without their consent.8 But this reform was implemented with doxxing, not deepfakes, in mind. It aims to prevent the unauthorised publication of an individual’s address or bank details as opposed to an AI-generated photo or video made in their image.
Australian politicians have instead turned to the criminal law for answers.
It has long been an offence under s474.17(1) of the Criminal Code to use a carriage service in a menacing, harassing or insulting manner. It comes with a jail term of five years, which increases to six if the offending involved the use of ‘private sexual material’. At first blush, these provisions appear broad enough to capture the creation and distribution of deepfakes. In fact, Australian Federal Police (AFP) have been using them for exactly that purpose.9
But the Commonwealth Director of Public Prosecutions (CDPP) is more circumspect. It questions whether an AI-generated photo or video constitutes ‘private sexual material’. In its submission to the Senate Standing Committee on Legal and Constitutional Affairs, which was debating whether to criminalise the creation and distribution of deepfakes, the CPDD opined: ‘It cannot be said that any expectation of privacy attaches to [an AI-generated] depiction of a victim’.10
Evidently, federal MPs agreed with the CDPP. Our laws were inadequate. After inviting submissions and holding public hearings, the Australian Parliament criminalised AI-generated, image-based abuse late last year. It’s now an offence to use a carriage service to transmit sexual material without consent, regardless of whether that material has been created or altered by technology. If convicted, a perpetrator faces a six-year term of imprisonment.
States and territories are following suit. Just this month, the New South Wales Government announced plans to amend the Crimes Act 1900 (NSW) to make the production of sexually explicit deepfake photos and videos, which depict real people, an offence carrying a three-year jail term. A similar Bill in South Australia would see perpetrators spend four years behind bars.
These reforms have deservedly made news headlines, but how will they be enforced? If the experience of the AFP tells us anything, it’s that police in New South Wales and South Australia have a difficult task ahead of them.
The union representing AFP officers has described investigating individuals who create and distribute deepfakes as ‘exceedingly challenging’.11 As AI technology becomes more sophisticated, police are struggling to determine whether the person depicted in an image or video is real or fake. Officers say the next step – locating a victim and the perpetrator – is equally arduous and time consuming. Anyone reading the union’s public submission to the Senate will have detected the exasperation of its members who posed this question: ‘How long do investigators spend trying to find a child who potentially doesn’t even exist or who had their likeness stolen but has ultimately not been abused themselves?’12
This question is revealing for two reasons. The first is that police undoubtedly require additional training and resources to enforce these new laws without compromising other investigations. The second is that police need education on the devastating impact of AI-generated, image-based abuse on its victims, which include both adults and children. They may not have been physically harmed, but – as one international study revealed last year – they’re more than twice as likely as non-victims to report psychological distress.13 Their distress is precisely why these laws were needed in the first place.
The ALA thanks David Lewis for this contribution.
For more reading on abuse law and trauma, please see the latest edition of Precedent available here (ALA members only).
The views and opinions expressed in this article are the authors and do not necessarily represent the views and opinions of the Australian Lawyers Alliance.
Learn how you can get involved and contribute an article.
1 E Rix, ‘Sydney high school senior investigated by police over deepfake pornographic images of female students’, ABC News (9 January 2025).
2 N Woodall, ‘Police investigate sexually explicit deepfake images of students at Gladstone Park Secondary College’, ABC News (21 February 2025).
3 eSafety Commissioner, Submission 28 to Senate Standing Committee on Legal and Constitutional Affairs Legislation Committee, Criminal Code Amendment (Deepfake Sexual Material) Bill (23 July 2024) 4.
4 K Sullivan, ‘South Australia set to criminalise AI deepfakes’, 9 News (30 April 2025).
5 Sections 16 and 162.
6 eSafety Commissioner, above note 3, 9.
7 Commonwealth, Parliamentary Debates, House of Representatives, 12 September 2024, (Mark Dreyfus, Attorney-General).
8 Privacy and Other Legislation Amendment Bill 2024 (Cth) sch 1, pt 3, s7.
9 AFP, Submission No 2 to Standing Committee on Legal and Constitutional Affairs, Criminal Code Amendment (Deepfake Sexual Material) Bill 2024 (19 July 2024) 4.
10 Director of Public Prosecutions (Cth) (CDPP) Submission 18 to the Senate Legal and Constitutional Affairs Committee, Criminal Code Amendment (Deepfake Sexual Material) Bill 2024 (19 July 2024) [9].
11 AFP, above note 9, 4.
12 Ibid.
13 eSafety Commissioner, above note 3, 5.