Queensland has introduced a bill that would make it a crime to create a sexual deepfake of a real person without their consent, even if the image or audio is never shared. The proposed offence carries a maximum of three years in prison.
Attorney-General Deb Frecklington announced the bill on 27 October, saying that “behind every deepfake image is a real person whose privacy, dignity and sense of safety has been violated”. The Criminal Code (Non-consensual Creation of Intimate Images and Audio Material) Amendment Bill 2026 must still pass the state parliament.
In brief
- The bill would create a new offence of making intimate images or audio of a real person without consent using digital tools, including AI and “nudify” apps.
- The maximum penalty would be three years; ABC News reports a lower maximum of 18 months for child offenders.
- Images that could be mistaken for real and identify a person are covered, and an AI-generated label may not take them outside the law.
- England and Wales made creating such images an offence in February, but with lighter penalties.
Queensland’s code already covers sharing, not making
Queensland’s Criminal Code already makes it an offence to record, distribute or threaten to distribute intimate images without consent. According to the bill’s explanatory notes, those offences fail to “adequately recognise or criminalise the distinct harm” of creating such material in the first place. The notes argue that once a fake exists, there is “a persistent risk” it will be misused.
The new offence would cover anyone who uses photo-editing software to alter an existing image, or directs AI software to generate intimate images or audio. The ministerial statement says it would include “nudify” applications and simulated people who “closely resemble, or purport to be, an identifiable real person”. Including audio means the offence would reach cloned voices as well as pictures; voice cloning is already familiar from phone scams.
The bill itself tightens the definition of an intimate image. A fake counts if it “could reasonably be mistaken for a genuine image” and identifies the person, and the explanatory notes say an image may still count even if it carries a disclaimer that it is AI-generated. Material so obviously fictional that no reasonable person could think it real, such as a cartoon, is excluded.
A child under 16 cannot consent to such material. Defendants could argue a genuine artistic, educational, legal, medical, scientific or public-benefit purpose, but would have to prove it. Police acting in the course of their duties are exempt.
The bill also lets courts order a person who has been charged, not only one convicted, to delete or destroy the images, with up to two years in prison for ignoring the order. Courts could bar convicted adults from contacting victims for up to five years.
The numbers behind the bill come from Australia’s eSafety Commissioner
The explanatory notes cite eSafety Commissioner figures showing sexually explicit deepfakes online rose by 550% between 2019 and 2023, that 98% of deepfake material was pornographic and that 99% of that imagery showed women and girls. In 2025, the notes say, the Commissioner found at least one incident of image-based abuse in Australian schools every week. Frecklington said teachers and students were “on the frontline of this emerging threat”.
ABC News reported that Frecklington pointed to the recent sentencing of Antonio Rotondo, who created and distributed sexualised deepfakes of current and former students and staff at Brisbane high schools.
Queensland’s proposed penalty is heavier than England’s
The closest UK comparison is section 66E of the Sexual Offences Act 2003, which made it an offence in England and Wales to create a “purported intimate image” of an adult without consent. It came into force on 6 February 2026 under the Data (Use and Access) Act 2025, alongside a separate offence of asking someone else to create one.
The two laws differ in three ways. Section 66E is a summary-only offence, punishable by up to the maximum term for summary offences or a fine; Queensland proposes up to three years. The English offence covers images that appear to be of an adult, while Queensland’s covers anyone and adds intimate audio. And England’s defence is any “reasonable excuse”, which the UK government must review within two years, whereas Queensland lists specific purposes.
That gap matters because the harm being targeted is the same. Regulators elsewhere are struggling with the platforms that host such tools: Ofcom’s chief executive told MPs the legal bar for an urgent court order over Grok deepfakes was too high, MLex reported, and Apple has updated its App Review rules on AI intimate imagery. Queensland’s approach puts the weight on the person who presses the button. If the bill passes, the offence will start 28 days after royal assent.
Sources
- Queensland Government: New offence targets AI deepfake sex abuse predators, 27 October 2026
- Queensland Legislation: Criminal Code (Non-consensual Creation of Intimate Images and Audio Material) Amendment Bill 2026, as introduced
- Queensland Legislation: Explanatory notes to the Criminal Code (Non-consensual Creation of Intimate Images and Audio Material) Amendment Bill 2026
- ABC News: Queensland government introduces new laws criminalising creation of sexually explicit deepfake images, 27 October 2026
- legislation.gov.uk: Sexual Offences Act 2003, section 66E
- legislation.gov.uk: The Data (Use and Access) Act 2025 (Commencement No. 5) Regulations 2026 (SI 2026/31)
Related stories
- Ofcom Weighed Court Order Over Grok Deepfakes but Bar Was Too High, MLex Says
- Apple Updates App Review Rules on AI Intimate Imagery and Bait-and-Switch
- 28% of UK Adults Think Voice-Clone Scammers Targeted Them in a Year
- Pennsylvania Signs AI Ad Disclosure Law With Fines Up to $3,000
Follow @tsnmediahub on X for breaking AI news as it happens.
Follow on X





