12 Days of ChristmAIs: A TMT insight series
AI is the latest and greatest toy on everyone’s wish list. It is transforming digital marketing by creating and automating advertising campaigns. At the press of a button, AI can deliver tailored customer experiences and provide targeted recommendations based on customer preferences and behaviours.
However, just like Christmas lunch, it’s easy to enjoy too much of a good thing. While AI can be an effective advertising tool, it poses compliance risks, particularly in relation to the advertisement of therapeutic goods in Australia.
In November this year the Therapeutic Goods Administration (TGA) provided updated guidance about using artificial intelligence (AI) to advertise therapeutic goods on social media, with the view to supporting improved compliance. We discuss the TGA’s new guidance below.
What is a ‘therapeutic good’?
The Therapeutic Goods Act 1989 (Cth) (TG Act) and the Therapeutic Goods Regulations 1990 (Cth) (TG Regulations), (together, the TG Rules) aim to control the quality, safety, efficacy and availability of therapeutic goods in Australia.
Therapeutic goods comprise a broad range of health-related products e.g. prescription medications, vitamins, herbal medicines, vaccines, medical devices, bandages, paracetamol and more. Food and cosmetics are generally not therapeutic goods; however cosmetics that make therapeutic claims or have a therapeutic purpose may be considered a therapeutic good e.g., moisturisers that contain a sunscreen agent as a secondary component could be classed as a therapeutic good and therefore subject to the TG Rules.
The Code
The Therapeutic Goods (Therapeutic Goods Advertising Code) Instrument 2021 (Cth) (the Code) outlines advertising and packaging requirements for the promotion of therapeutic goods in Australia. Failure to comply with the TG Rules and Code may result in the issuance of infringement notices, directions or prevention notices, and civil or criminal penalties.
Advertisements must (among other things):
- be accurate, balanced and not misleading;
- support the safe and proper use of the therapeutic goods;
- not cause undue alarm, fear or distress; and
- be reviewed to ensure the content of testimonials and endorsements are verified.
Using AI to advertise therapeutic goods
In its November update, the TGA acknowledged the rise of AI technology and its use in the digital marketing environment to create, automate and deliver personalised experiences and drive website traffic.
However, in circumstances where business owners and advertisers have engaged a third-party AI service provider or tool to generate advertising content, the TGA has confirmed that the responsibility to ensure AI-generated advertising content complies with the TG Rules and Code rests with business owners and advertisers. The responsibility covers influencer endorsements and user-generated content including third party comments posted on social media platforms.
Using AI-generated advertisements or AI-assisted advertising content is not unlawful. However, all AI outputs should undergo robust and adequate oversight by a human.
Business owners and advertisers must ensure that all advertising content, including materials generated by, or with the assistance of AI, complies with the TG Rules and Code. This includes both current and historical posts and content. Accountability rests with business owners and advertisers to oversee any AI outputs regardless of how big or small their involvement.
Risks associated with using AI in advertising
Using AI in business can create efficiencies and keep costs down. However, AI-generated advertising content poses non-compliance risks, including:
Dissemination of inaccurate and misleading information
In generating its outputs, AI models tend to prioritise engagement, not accuracy. If product packaging incorporates a particular claim, AI may generate advertising that repeats or amplifies non-compliant language to increase consumer engagement.
Contrary to the actual characteristics of goods, AI is likely to prioritise consumer desires and might use terms such as ‘fun’, ‘natural’, or ‘harm-less’, or omit ingredients within promotional content to persuade consumers to purchase the goods, which can lead to non-compliance with the Code.
Automatic production and modification of data not substantiated by the advertiser
Once again, AI prioritises engagement over accuracy and often uses click-bait hooks to gain audience attention. For example, describing goods as ‘new and improved’ or the ‘best on the market’, when these claims might be factually incorrect.
Making claims that cannot be substantiated through evidence is a breach of the Code. It is important to verify the claims made in AI advertising campaigns before publication or dissemination.
False or misleading reviews and testimonials
Business owners and advertisers should ensure any use of AI does not create synthetic media or misleading representations of real people, false endorsements, or misused celebrity likeness (such as deep fakes), without permission and verification. Advertisers should not blindly adopt AI generated endorsements, or any other comments, without first ensuring compliance with the Code.
False or misleading AI chatbot referrals
As a result of recent changes to how popular search engines generate and display search results (e.g. through an AI summary at the top of search results), the market is observing a trend of business owners relying on referrals through popular AI chatbots (e.g. ChatGPT).
However, the use of AI chatbot referrals or endorsements also creates advertising compliance risks. This is because AI often exaggerates the usefulness of products and may not base comments on a genuine, unbiased account of an ordinary consumer’s use of the product.
Business owners and advertisers may have little control or oversight over AI chatbot output. Yet it is business and advertisers owners who are responsible for compliance with the TGA Rules and the Code.
Walking in an AI wonderland
The rise and use of AI is here to stay. Using AI-generated advertisements or AI-assisted advertising content is not unlawful. However, all AI outputs should undergo robust and adequate oversight by a human. Otherwise, there is a real risk the packaging and advertising of therapeutic goods may be deemed non-compliant under the TGA Rules and Code, leading to the issuance of infringement notices, directions or prevention notices, and civil or criminal penalties.
In its update, the TGA recommend businesses publicly provide corrective information if they become aware of the publication of any misinformation about their products.
To regulate social media comments, the TGA has also suggested that business owners and advertisers develop a social media acceptable use policy which sets out guidelines to comply with their obligations under the Code.
If you have any queries or require any assistance regarding AI-supported advertising practices, please reach out to expert Digital & Intellectual Property team. We are here to help you navigate this evolving regulatory landscape with confidence.
