Whether predicting personalized content and customer behavior, using it by agencies to optimize their advertising campaigns, or training brands in large proprietary language models tailored to their specific brand identity, l he use of AI in marketing is booming. So much so that the global AI in marketing market size is expected to reach $72.1 billion by 2030, a six-fold increase from 2022.
As always, few booms are exploited without financial and reputational risk. Under Armour, for example, made a splash earlier this year when it unveiled an ad featuring British boxer Anthony Joshua in a fast-cut black-and-white montage using recycled footage made for previous ads.
Not surprisingly, critics have questioned the ethics of reusing old work. Then there are the many brands who worry about mistakenly publishing work that infringes copyright, or who worry that feeding their clients’ information into an AI system could help train the model of a competitor, for example.
The risks are such that marketers have started adding clauses to agency contracts to prohibit the use of any form of AI without prior approval. To explore these types of challenges, the Advertising Association aptly launched an AI taskforce last autumn aimed at helping the UK industry deal with both the promise and perils of AI.
I wonder, however, whether there is enough focus on how marketers can make the most of what AI has to offer while avoiding the gender biases and distortions inherent in AI. Biases that can and do distort AI-based insights and recommendations for marketers. And if we don’t move fast enough and work hard enough to disfavor AI and the way it’s deployed in marketing – and it is indeed hard work – do we risk seeing some of the recent progress of our industry on how women are marketed to come undone. ?
The problem is most immediately visible when it comes to generating visual content. Many AI image generation models are trained on datasets scraped from the internet, which often perpetuate unrealistic and stereotypical depictions of women. “These data sets often overrepresent traditionally attractive young women and depict them in sexualized or subordinate roles. As a result, when marketers use these AI tools to create visual content, they may inadvertently reinforce harmful gender stereotypes,” Rhonda Hadi, associate professor of marketing at the Saïd Business School of the University of Oxford. (Hats off here to Dove and the brand’s recent commitment to not use AI-generated images to depict women in its advertising and communications).
AI as a tool can both compound existing gender biases and also help fill crucial data gaps resulting from gender discrimination.
“The problem of data bias is real,” adds Candina Weston, AI specialist, consultant and former CMO at Microsoft. “The first step to leveraging AI technology at scale must be to ensure the work is done to bring the target audience’s data to the right level to achieve the right outcome. This is an ongoing process, never a one-off, and is true with or without AI.
And that’s a challenge because if, like me, you think that too much data purported to be representative of women today is at best incomplete and at worst reductive, then this suggests that much of the use of AI in our industry is going to be irrelevant from the start. Ironically, perhaps, a solution to rectifying the situation in the face of non-existent data (for example in the financial sector – remember that it was not until 1975 that British women were able to open a bank account in their own name , which therefore represents a whole bunch of important historical facts). missing data on women’s creditworthiness) is synthetic data – data that is artificially generated by algorithms or simulations and can be used to train machine learning models.
Clearly, AI as a tool can both worsen existing gender biases and also help fill crucial data gaps resulting from gender discrimination – a stark reminder that it does not There is nothing black or white in the arguments for or against AI in marketing. Like many of her peers, Tamara Rogers, global marketing director at Haleon, believes the relationship between marketing and AI should be a delicate dance between ambition and caution.
For example, as her teams have used AI tools like CreativeX to improve marketing measurement and analysis at scale, across hundreds of creative assets, she says: “This rich level of insight allows us helps ensure that our marketing and advertising investments are targeted. on the best areas and engages and resonates with people in the right place, at the right time. A benefit that clearly reflects Haleon’s ambition to drive greater health inclusion globally: there are limits to the parts of the campaign trail for which Haleon will deploy AI. “It is difficult, for example, for AI to replicate the generation of original ideas, an essential part of the creativity that human minds and skills can bring when designing content, especially when it is necessary reflect local cultural contexts and other nuances within the countryside,” says Rogers.
Until data sets are completed and corrected, the ultimate guardrail that marketers and their agencies must integrate and elevate is the last-mile audit process.
So, reading this that generative AI is not yet quite the large-scale threat that some creative agencies may have feared, there is an argument that our industry should spend less time obsessing over something something that remains existential for now and more time to have essential conversations about the real bias in the data that currently powers many marketing AI tools. Not only because of the moral imperative, but also because if we can solve the problem, we are on the verge of achieving the holy grail of our industry: the ability to create tailor-made marketing for individuals at the same time. global scale.
“Improving data will allow us to get closer to hyper-personalization. I think it’s something that isn’t talked about enough, but is so critical to allowing us to stop unwanted elements and access the useful things,” says Weston.
And let’s face it, as a historically underserved and overlooked target, as a working mom with a mental load heavier than those unused dumbbells in my bedroom closet, I wouldn’t mind a little hyper marketing personalized and USEFUL comes my way.
Until data sets are completed and corrected, the ultimate guardrail that marketers and their agencies must integrate and elevate is this last-mile audit process, a level of human review for everything that goes directly to a customer (assuming those humans don’t). nor does it carry the same bias). I would also like to think that, as Dove illustrates, we are evolved enough – as an industry – to know when to limit the types of areas in which we choose to deploy AI for marketing campaigns. We must not let the seductive siren call of AI lure the industry back in time, undoing its hard (and ongoing) work to bias and dispel stereotypes in everything we do and produce.
