A stunning model wearing a glamorous white dress to promote Melbourne fashion house Atoir’s $400 item on The Iconic is one of the latest examples of how AI in marketing can cause reputational damage.
The problem here was that neither the dress nor the model was real – both were AI-generated, causing eagle-eyed online shoppers to call out the brand for deceptive behaviour as there was no way the image replicated how the garment sits on the body.
This is a good reminder for all organisations, and particularly for marketing and communications professionals, of the importance of having clear controls in place to avoid the same, or much worse, reputational damage as Atoir and others have experienced.
It is well acknowledged that AI is being adopted at breakneck speed by organisations across the globe, including in Australia, with Deloitte data outlined in the ACS Digital Pulse 2024 report showing that 54 per cent of students and employees use gen AI for any purpose.
AI Usage in Marketing
Marketing is the department in any organisation most likely to adopt AI tools, which are revolutionising everything from customer segmentation, competitor analysis, and content generation through to personalisation and predictive analysis.
While there are some avowed AI naysayers – due to concerns around loss of human values, operational risks, job security and misinformation – data by global research firms suggests that 60 to 70 per cent of those employed in marketing roles use AI in some way.
Further, a BCG survey released in June this year, found 71 per cent of more than 200 CMOs reported that they planned to Invest more than $10 Million annually on Generative AI in the Next Three Years.
And for good reason.
A Marketing Game Changer
As reported in the BCG report, launched at Cannes, benefits identified by utilising AI included improvements in customer experience, content quality and volume, efficiency gains from increased productivity and the elimination of manual work.
AI powered applications have been used to boost campaign efficiency, improve customer engagement, and even produce TVCs faster, more efficient and at significant cost savings such as Amaysim’s “The Escape Story” fully AI-Generated TVC delivered by a two-person in-house creative team.
Other examples of AI playing a positive role in marketing include:
- hyper-personalised messages we receive from Netflix and Spotify recommending new content based on viewing/listening behaviours.
- predictive analysis, with the NAB’s “Customer Brian” setting the standard by anticipating individual needs across products and interactions, resulting in an increase in customer engagement of around 40 percent.
- Optimisation of advertising bids, targeting and creative of programmatic advertising campaigns.
- and assistance (note the word assistance) in creating content such as product information, blogs, and social media at scale.
The Risk of Losing Control
However, over-reliance on AI in marketing and communications can lead to embedded risks without appropriate oversight and governance.
Atoir is not the only example of an AI misfire, resulting in shaken consumer trust, as Microsoft and LG Hooker can attest from their use of AI to produce advertising content.
In the case of Microsoft, a LinkedIn job ad for Xbox Australia, which included an AI-generated image of a person coding, with code displayed on the back of the monitor, sparked online outrage – particularly given the company announced significant cuts to its creative and technical roles only weeks earlier. In this case, it resulted in reputational damage, was inconsistent with the brand tone and voice and eroded trust.
For LJ Hooker, in addition to the same negative outcomes experienced by Xbox, it also created some legal and compliance issues, with the ACCC noting that under “consumer law, businesses should not engage in false or misleading conduct in the supply of goods or services, this includes in advertising properties for sale or rent”.
But the risk is not just from what goes out to the world, but also what is put into the AI machine.
This a critical bling sport for many companies as they are giving their people free reign to use platforms as they like including the use of the free versions made available to help train the platform.
This unfortunately means that information fed into the free version of the platform, no matter how confidential or sensitive it is, becomes part of the platforms learning and as such could be echoed back as fact.
Samsung is one business that found this out when in 2023 it became evident that, in three separate incidences, staff had fed sensitive code and proprietary details into ChatGPT’s free version and as a result this information was inadvertently leaked.
The Bottom Line
While the adoption levels of AI varies between industries and companies, and how it is controlled differs in many ways, AI is here to stay.
It is not a silver bullet for all things marketing – as you would appreciate if you ever read the content it produces that feels robotic or impersonal and lacks nuance, voice or warmth typically found in human writing.
But implemented in the right way with governance and upskilling, it can improve efficiencies, customer engagement and innovation.
Author’s note: In writing this article, our enterprise-grade Chat GPT account was utilised to help research some data points and details of the case studies I recalled reading about over the past few years (which I then verified via other sources to ensure the output was correct). It also suggested some headlines for the article following my prompting, though have opted against using any of the recommendations, given the bias they had towards AI. Finally, another AI-powered engine, Microsoft Editor gave a quick review of this piece, resulting in a few minor edits that I had not picked up on.
