Introduction
Artificial intelligence (‘AI’) has begun to dominate the globe with deepfake technologies becoming a ubiquitous tool for digital users. With an abundance of resources at the fingertips of the user, it is easy for the unsuspecting eye to be deceived by fake imagery created by artificial intelligence. The eSafety Commissioner of Australia describes a ‘deepfake’ as: ‘…a digital photo, video or sound file of a real person that has been created with AI (artificial intelligence) to make an extremely realists but false depiction of them doing or saying something that they did not actually do or say.’
With this being a growing contention, especially within the legal sector, questions arise as to what are the implications of deepfakes affecting the law and how is the law in Australia tackling the issue? We discuss the current implications facing Australian law, as well as the current challenges that are being presented to the Australian courts.
Deepfake Technology Within the Course of Legal Practice
Due to deepfake technologies becoming a user-friendly platform for most users, the issue is of upmost importance for the Australian lawyer to be informed on all fronts as to how to adapt and control the use of deepfake technology in their everyday practise and to their fellow colleagues within the legal profession.
When referring to the state regulations of the Australian Solicitors’ Conduct Rules 2015 (NSW) and the Legal Profession Uniform Conduct (Barristers) 2015 (NSW), both regulations stipulate, that it is the duty of the solicitor and barrister to ‘maintain high standards of professional conduct’ amongst their duty to their clients and to the court, and to always be ‘honest and courteous’ in all of their dealings during the course of their legal practice. However, it is easy for the Australian lawyer to take deepfake technology for granted when working on complex legal tasks without giving pieces of information a second glance.
The prevention of using such evidence that is deemed ‘fraud’ or a piece of information that is ‘hallucinate’ from AI is paramount.
Current Legal Position of Deepfakes in Australia
As of the publishing of this article, Australia is yet to introduce a standalone comprehensive Act into law. Pieces of legislation regarding the use of AI, specifically deepfake imagery, however, can be found within the Privacy Law Act 1988 (Cth).
In addition to existing legal provisions, the ‘Australian AI Ethics Principles’ were introduced in 2019 with its core purpose in being to guide ‘business and government to responsibly design, develop and implement AI.’ Additionally, these guided principles were designed to achieve an environment where AI is safe, reliable and fair to use for Australians, reduce the risk imposed when using AI, and to maintain a high ethical standard when using AI.
The key factors of the AI Ethics Principles include, but are not limited to:
- Human, Social and Environmental Wellbeing
1.1. This principle states that AI should be used for beneficial purposes. - Human-Centred Values
2.1. This principle states the importance of AI being used to ‘serve humans, and not the other way around.’ - Reliability and Safety
3.1. This principle states that AI should not be used to pose any unreasonable safety risks to users, and that ‘AI systems should be monitored and tested to ensure they continue to meet their intended purpose…’ - Privacy Protection and Security
4.1. This principle states that AI must ‘ensure respect for privacy and date protection when using AI systems,’ and ‘to ensure appropriate data and AI system security measures are in place.’
Current Legal Position in International Law
Compare this with our European counterparts where the European Union has introduced ‘The EU Artificial Intelligence Act.’ This legislation’s opening text addresses the following:
The purpose of this Regulation is to improve the functioning of the internal market by laying down a uniform legal framework in particular for the development, the placing on the market, the putting into service and the use of artificial intelligence systems (AI systems) in the Union, in accordance with Union values, to promote the uptake of human centric and trustworthy artificial intelligence (AI) while ensuring a high level of protection of health, safety, fundamental rights as enshrined in the Charter of Fundamental Rights of the European Union (the ‘Charter’), including democracy, the rule of law and environmental protection, to protect against the harmful effects of AI systems in the Union, and to support innovation. This Regulation ensures the free movement, cross-border, of AI-based goods and services, thus preventing Member States from imposing restrictions on the development, marketing and use of AI systems, unless explicitly authorised by this Regulation.
Furthermore, deepfakes are recognised under the Act whereby a user that deliberately generates and/or manipulates an image, an audio clip or a video, should be disclosed to not deceive the public.
The Tort of Deepfakes
While Australia does not have a standalone legislation that addresses the issue of Deepfake technologies directly, this has not stopped the common law from being aware of its implications and its effect on the Australian public.
In the Victorian County Court case of Waller (A Pseudonym) v Barrett (A Pseudonym), her honours Justice Tran made the remark that ‘…there is an increasingly urgent need to protect privacy in the world of […] deepfakes […].’ Further, Tran J addresses how deepfakes contributes to the issue of privacy in the following:
The world in which we now find ourselves is one of deepfakes, including deepfake pornography; a decline in market share of big media; and the pervasive, “democratising” force of social media. To establish their privacy has been invaded, is a plaintiff to be required to prove that a sex tape is “real”? Must a plaintiff suing because a viral social media post is written concerning their sexual preferences, first establish, by admissible evidence, their sexuality? If the social media post is false, or the video a fake, is a plaintiff’s only recourse the law of defamation? What if the matters disclosed are intensely private, but not defamatory? If the action is so limited, an aggrieved person would effectively be placed into a catch 22: to obtain relief they would need to establish the truth of matters said about them, thereby exposing their intimate life even further. Meanwhile, the defendant who makes inaccurate public statements about intensely private matters is in a better legal position (and probably has more “clicks”) than the defendant who properly fact checks and makes only accurate statements.
In a latter case heard in the Federal Court, their honour Longbottom J ruled in the case of eSafety Commissioner v Rotondo that the respondent had contravened s 75 of the Online Safety Act 2021 (Cth) by posting a deepfake ‘moving visual image’ on a website which depicted another ‘person’s genital area and breasts.’ It had also depicted the other person engaging in a sexual act without their consent. s 75 provides that a person ‘must not post, or make a threat to post, an intimate image of another person.’
Their Honour was satisfactory to establish that the Respondent’s actions of posting the non-consensual deepfake imagery had caused the victims hurt and distress. Supporting their reasons, their honour notes regardless of how “real” or “deepfake” a sexually explicit image may be, sharing such content “feel[s] real” for victims.
Criminal Intention
To address the severity of deepfake imagery in Australia, the Government has introduced an amendment to the Criminal Code Act 1995 (Cth), declaring that the creation of ‘deepfake sexual material’ as a criminal offence. s 474.17A states that the sharing of deepfake sexual material without consent is an offence if the person shares material of another person that the other person does not consent to, and that material depicts the other person ‘engaging in a sexual pose or sexual activity.’
The recent amendment was put to the test in the recent Adelaide Magistrates Court case of William Hamish Yeates. William had pleaded guilty under the new federal provision; guilty to two counts ‘of creating or altering sexual material without consent and two counts of using a carriage service in a harassing or offensive. way’ The accused had previously been charged with 20 offences by the Commonwealth; however, most of these charges were dropped after admitting to the creation of the non-consensual sexual material.
Conclusion
The concept of harmonising current AI provisions into a singular piece of legislation can be viewed as a challenging road ahead for lawmakers. With no cohesive or robust piece of legislation that is designed to combat AI and specifically deepfakes, Australia may fall behind with the rest of the world in regulating the growing concern of using deepfake technologies. The common law, however, does provide insight for lawmakers to analyse current issues that are affecting the legal landscape of using such AI technologies in Australian law as was recognised in the recent cases discussed. Deepfakes pose a global risk to the landscape of law, and without a sustainable legal framework to regulate, and monitor deepfake technologies, deepfakes will only continue to pose more harm than it does good.
