An online abuse expert and Canterbury University senior law lecturer Dr Cassandra Mudgway said there was little individuals could do to protect their images from being used as deepfakes – other than not posting them online.
Mudgway told Checkpoint the tools that could nudify an image were easy to use and easy to access.
“You just upload it, you press a button, you put in a prompt and the AI tool gives you the result.”
She said unlike some other countries, New Zealand had not banned apps, websites or tools that allowed nudifying images.
“I think that we could follow the Australian example and think about banning them, basically geo-blocking them so you cannot access them.”
Mudgway said currently nudifying tools could also be accessed on mainstream platforms such as Google and Apple’s app stores.
“Those terms and services say that there shouldn’t be any apps for sale that are expressly for sexualised deepfakes, but that doesn’t mean that they’re not there. They’re often disguised as face swap app, so they’re still being sold there.”
She said there were also incidental tools that were not primarily for sexualised deepfakes but could still be used that way.
“Think for example Grok by Elon Musk.”
The Deepfake Digital Harm and Exploitation Bill is making its way through Parliament and would amend the Crimes Act and Harmful Digital Communications Act by expanding the definition of ‘intimate visual recording’ to include images created, synthesised or altered to show a person’s likeness in intimate contexts without their consent.
Mudgway said this would be of most benefit to victim-survivors who were adults.
“It’s going to be an avenue to get justice for individuals, but it’s not going to tackle the scale of what we are dealing with here.”
She said in most cases if sexualised deepfakes had victims under 18, that was already considered a crime under the Films, Videos and Publications Classification Act.
Alongside wider regulation around AI tools, Mudgway also an opportunity to educate young people on the harm of AI tools and encourage AI scepticism.
