ACT List MP Laura McClure has told a parliamentary select committee her member's bill targeting sexually explicit deepfakes is a "first step," urging MPs to look beyond it toward wider law reform, education, and possible regulation of AI.

McClure, the ACT Party's Education Spokesperson, appeared before the Social Services and Community Committee on a Wednesday to discuss the Deepfake Digital Harm and Exploitation Bill. The bill passed its first reading unanimously on 20 May 2026 and was sent to the committee for public submissions, which closed on 19 June 2026 (Parliament).

A deepfake is a fake image or video created using artificial intelligence — in this case, sexually explicit material made to look like a real person without their consent. The bill would change the Crimes Act 1961 and the Harmful Digital Communications Act 2015 by expanding the legal definition of an "intimate visual recording" to include AI-manipulated images or videos made without consent. That would make creating, sharing, or selling sexually explicit deepfakes illegal. McClure has said the bill would make a nude AI-generated fake image as illegal as a real non-consensual nude image (ACT Party).

She told the committee that police struggle to prosecute explicit deepfake cases under existing law because they must prove the person intended to cause harm. McClure said a successful prosecution earlier in the year had required the victim's family to build an extensive case themselves, pointing to the gap her bill seeks to close (RNZ).

McClure also told MPs the Harmful Digital Communications Act, passed in 2015, needed a full review. She suggested a "national hashing register" as one option worth considering. Hashing works like a digital fingerprint for a file — once an image is registered, platforms can use that fingerprint to detect and block the same image being shared again. She said deepfaking could be used for fraudulent purposes beyond pornography but described that as a "massive piece of work" for a future government (RNZ).

The bill was drawn from the member's bill ballot and has attracted cross-party support. Te Pāti Māori gave it formal backing ahead of the first reading (ACT Party). McClure has framed the legislation around the principles of personal autonomy and self-ownership (ACT Party).

The broader context here is that AI-generated content is moving faster than the laws designed to manage it. The Harmful Digital Communications Act is over a decade old and was written before generative AI tools became widely available. The intent-to-harm threshold McClure identified has been a practical barrier: police must show that someone who created or shared a deepfake meant to cause harm, which is hard to prove when material spreads through peer groups or encrypted channels where motive is difficult to pin down.

The hashing register idea McClure floated is not new in concept. PhotoDNA-style hash-matching technology is already used by major platforms to detect and remove known child sexual abuse material. Extending that model to non-consensual intimate imagery, including deepfakes, would need infrastructure, coordination between platforms and government, and a way for victims or police to register hashes of specific images. McClure's suggestion was offered as an option worth considering rather than a formal proposal within the bill itself.

The bill's narrow focus on sexually explicit deepfakes, while leaving broader fraudulent uses for future work, reflects a common pattern with member's bills: keeping the legislation narrow enough to hold the cross-party support needed to pass under MMP. The unanimous first reading signals broad agreement on the principle that non-consensual sexual deepfakes should be illegal. The select committee stage is where that consensus gets tested against the specifics of how the law is drafted and what it requires of platforms, police, and individuals.

McClure's call for wider reform, more education, and possible AI regulation goes well beyond what her bill does. It positions the legislation as a floor, not a ceiling, on Parliament's response to AI-generated harm. Whether a future government picks up that broader work is an open question.

The Social Services and Community Committee will now consider the public submissions received before the 19 June deadline and report back to the House. The bill's passage through the remaining stages will depend on whether the unanimous support at first reading holds once members have debated amendments and the committee's recommendations.