Civil Law

By Caitlin Holloway

07.02.24

MP Georgie Purcell’s AI breast ‘enhancement’ leaves Australia’s legal system exposed

Victorian upper house MP Georgie Purcell last week slammed Nine News for using an image of her that had been edited to make her breasts look bigger and expose her midriff.

Nine News apologised to Ms Purcell, claiming it was a “graphic error” caused by the use of “automation by Photoshop”, a generative artificial intelligence (AI) tool available in Adobe Photoshop.

While a spokesperson for Adobe has said the use of its generative AI features would have required a level of human intervention, the publication of the edited image of Ms Purcell raises questions over how our legal system can adequately address the increasing use of AI, and support individuals like Ms Purcell.

As it stands, existing laws around defamation and copyright are left wanting when it comes to protecting Australians impacted by AI-generated content.

Many editing and design products, such as Adobe Photoshop, have recently integrated generative artificial intelligence tools. The use of AI in the creation or editing of content is becoming increasingly accessible to everyone – from regular individuals posting content online, to large-scale media corporations such as Nine News.

Why should we be concerned? I want you to imagine that a photo or video of yourself has been posted online, capturing you doing something unfavourable, or even illegal, that you didn’t do. Would the Australian legal system be able to step in and offer you a remedy? The answer is concerningly complicated and highlights a need for our legal system to catch up with advancements in this space.

Let’s have a look at two areas of law which are designed to step in when it comes to issues concerning the publication of content – defamation and copyright law.

Defamation law

Defamation is concerned with injury to a person’s reputation without reason or justification. In essence, it protects the value of your reputation and recognises that a person’s reputation can be harmed in a similar way to physical harm (albeit intangible).

To be successful in a claim for defamation, you must establish that a defendant published something to a third person or parties that:

  1. identifies you;
  2. contains “defamatory imputations” (or defamatory meanings); and
  3. has caused you serious harm.

Sounds easy, right? Well, not necessarily in the case of an image or video that is the result of generative artificial intelligence.

Traditionally, we see defamation claims that are the result of spoken or written words – news articles that impugn the reputation of an individual, or even posts on social media (like the example in one of our previous blogs: https://aulich.com.au/how-harmful-are-your-words/).

The use of defamation law to protect reputational damage caused by generative artificial intelligence is not as clear cut. A video or photo published online can clearly identify a plaintiff and cause serious harm to that person’s reputation but proving that the photo or video contained defamatory imputations is difficult when those imputations are not conveyed in spoken or written word.

Take Ms Purcell for example – what does the enhancement of her breasts and the exposing of her midriff mean for the purposes of defamation law? Was it the intention of Nine News to portray her as a woman not suited to politics? What did the people to whom the image was published think about Ms Purcell’s reputation when they saw the image?

The answer to those questions is unknown and highlights the barriers to using defamation law for a remedy in a case involving an image created by generative artificial intelligence.

Copyright law

Copyright law, on the other hand, is designed to protect original works. It provides creators with an exclusive right to their intellectual property and prevents others from reproducing or distributing works without permission.

There are a variety of remedies available to creators under copyright law. These can include injunctions (for example, a court order requiring the removal of the content in question) and damages (such as an account of profits made by the person who has reproduced the content).

While these remedies could be helpful in addressing the use of generative AI they also contain significant limitations.

To access copyright law remedies, you must be the owner of that intellectual property. In some cases this is clear cut, but in others it might not be. For example, an image of you could be superimposed in a video that has been created by someone else online. In that circumstance, you may not even have standing to bring a claim for breach of copyright because the owner of the video would be the person with intellectual property law remedies available to them.

Additionally, and unlike defamation law, copyright law is not designed to remedy harm to reputation – in fact, it does not even recognise it. Its purpose is to protect the “originality” of works.

These issues highlight the need for innovative change in our legal system, to adequately address some of the dangers of generative artificial intelligence.

At Aulich, our lawyers are experienced in navigating the complexities of defamation and intellectual property law, and are here to help if you’ve been the subject of defamation or a breach of your intellectual property law rights.