Understanding Deep Fakes

What is a deep fake?

A deep fake is a highly realistic but fabricated image or video where an individual's likeness is manipulated using artificial intelligence (AI).

These creations, facilitated by advanced technology, were once hard to make but are now easily accessible to the public through free downloadable apps on smartphones.

Are deep fakes common?

Deep fake technology has been part of our lives for years, especially in movie special effects, such as 'de-aging' actors. Initially, these effects mostly targeted public figures due to the extensive data required.

However, as technology rapidly advances, it's becoming easier to create realistic deep fakes of ordinary people, even with minimal photo or video references.

Are deep fakes harmful?

Whilst not inherently malicious, deep fakes pose significant risks due to their potential for misuse. The combination of sophisticated technology and accessibility makes deep fakes a powerful tool for exploitation. They are increasingly used for harmful purposes such as fake news(/online-safety-at-home/mis-disinformation), scams(/scams) and image-based abuse(/online-abuse-and-harassment/image-based-abuse) (historically known as 'revenge porn').

Global evidence shows a rising trend in women being targeted by these harmful practices. The personal harm caused by pornographic deep fakes, and fake nude images, can be severe and the non-realistic nature of these images does not lessen the damage to the targeted individuals. Manipulated videos of political figures are also on the rise, jeopardising democratic processes.

How to spot deep fakes

Detecting deep fakes is becoming more challenging as technology progresses.

Indicators to watch out for include flickering around the face and hair, inconsistencies in skin texture, unnatural eye movements or finger placements, and unrealistic behaviours or speech.

What can I do if someone has made a deep fake of me or someone I know

In New Zealand, harmful digital communications are actionable, whether real or manipulated, under the Harmful Digital Communications Act (HDCA)(/our-work/helpline-services/the-harmful-digital-communications-act).

Netsafe is the Approved Agency for the HDCA and our helpline handles confidential reports of such harms and supports you through the options available to you.

Deep fakes and the law in New Zealand

Despite worries that AI-created or modified images may not be legally covered, existing New Zealand legislation offers protections:

Harmful Digital Communications Act 2015: This Act addresses potentially harmful digital communications that may cause serious emotional distress, irrespective of their creation method. This includes AI-generated or modified images if they cause or are likely to cause harm.
Films, Videos, and Publications Classification Act 1993: A publication is deemed objectionable and illegal if it involves sex, horror, crime, cruelty, or violence in a way injurious to the public good. AI-generated or AI-modified images falling under this category are illegal.

Given the rapid pace of technological change, some laws may need updates to explicitly cover deep fake and AI-generated content.

Netsafe is advocating for the criminal offence of “posting an intimate visual recording without consent” to be amended to clearly include AI-generated content.

What should I do if I find a deep fake?

Before sharing a suspected deep fake, consider its intent and report it to the platform for review.

Sharing such content—even to raise awareness—can worsen the harm for those depicted. Social media platforms have policies to manage and mark deep fake content to minimise the harm caused by sharing it.