In marketing, from personalized content and predicting customer behavior, to agencies using it to optimize advertising campaigns, or brands training their own large-scale language models suited to specific brand identities. The use of AI is booming. Therefore, the global market size for AI in marketing is expected to reach $72.1 billion by 2030, a six-fold increase compared to 2022.
As always, few can take advantage of the boom without financial and reputational risk. Under Armor, for example, caused a stir earlier this year when it released an ad featuring British boxer Anthony Joshua in a black-and-white, fast-cut montage that recycled footage from previous commercials.
Naturally, critics questioned the ethics of reusing old works. Additionally, many brands worry about accidentally publishing copyright-infringing work or worry that inputting customer information into an AI system could help train a competitor's model. I am.
This risk is significant enough that marketers are starting to add clauses to agency contracts that prohibit the use of any type of AI without prior permission. To explore these kinds of challenges, the Advertising Association wisely launched an AI Taskforce last autumn, aimed at helping the UK industry navigate both the opportunities and dangers of AI. Ta.
However, the question is whether there is enough focus on how marketers can make the most of the capabilities AI offers while avoiding the gender biases and distortions inherent in AI. Biases that can skew her AI-driven insights and recommendations for marketers. And if we're not moving fast enough and hard enough to remove bias in AI and how it's implemented in marketing, which is certainly a daunting task, the industry's recent advances in women's marketing Is there a risk that it will be ruined? ?
This problem is most immediately apparent when it comes to visual content generation. Many AI image generation models are trained on datasets collected from the internet, often perpetuating unrealistic and stereotypical depictions of women. “These datasets often overrepresent young, traditionally attractive women, portraying them in sexualized or subordinate roles. As a result, marketers When they use these AI tools to create visual content, they can inadvertently reinforce harmful gender stereotypes,” Rhonda Hadi, associate professor of marketing at Oxford University's Saïd Business School, told me. (Kudos to Dove and the brand's recent pledge not to use AI-generated images to represent women in advertising and communications).
As a tool, AI can not only deepen existing gender biases, but also help fill important data gaps that arise from gender discrimination.
“The problem of data bias is real,” adds Candina Weston, an AI specialist, consultant, and former CMO at Microsoft. “The first step in leveraging AI technology at scale is ensuring that you are working to get your target user data to the right level and getting the right output. This is an ongoing process, and once you It's not a one-and-done kind of thing, and it applies with or without AI.”
And this is a difficult question. Because if, like me, you think that too much data that is supposed to be representative of women today is incomplete at best and simplistic at worst, then it's important to understand that AI is being used in our industry. Because it shows that much of what is being said is going to be off the mark from the start. Ironically, data that does not exist (for example, in the case of the financial sector, it was not until 1975 that British women could open bank accounts in their own names, so there is no significant history of women's creditworthiness). One solution for course correction (which results in large amounts of missing data) is synthetic data. Synthetic data is artificially generated through algorithms or simulations and can be used to train machine learning models.
It is therefore clear that AI as a tool can not only deepen existing gender biases, but also help fill critical data gaps arising from gender discrimination. This is a stark reminder that there is nothing black and white in the debate for and against AI in marketing. Tamara Rogers, Haleon's global CMO, believes, like many of her colleagues, that the relationship between marketing and AI should be a delicate dance between ambition and caution.
For example, she says, her team has used AI tools like CreativeX to drive large-scale marketing evaluation and analysis across hundreds of creative assets. “This rich level of insight allows us to ensure we focus our marketing and advertising investments in the right areas, at the right time, and in the right places to engage and resonate with people.” Masu. ”This is a benefit that underlines Hareon's ambition to drive improved health and inclusion around the world. There are limits to some of the campaigns in which Hareon deploys her AI. “For example, when designing content, especially when campaigns need to reflect local cultural context and other nuances, AI can help with original idea generation, which is at the core of the creativity that human thinking and skill sets provide. It’s difficult to replicate,” says Rogers.
The final guardrail that marketers and their agencies must build and strengthen until the data set is complete and unbiased is the last-mile audit process.
This shows that generative AI is not yet the large-scale threat that some creative agencies feared, and that our industry should spend less time fixating on what has survived for now and focusing more on the essentials. There is an argument that we should spend more time on social conversations. There are real biases in the data currently feeding many of marketing's AI tools. Not just because we have a moral obligation to do so, but because solving the problem brings us one step closer to the industry's holy grail: creating bespoke marketing for individuals on a global scale.
“Improving your data will bring you closer to hyper-personalization, which I don't think is talked about enough, but it's really important so that you can stop what you don't need and access what's useful. ” says Weston.
And as part of a historically underserved and ignored target population, as a working mother with a greater mental burden than a dumbbell lying unused in a bedroom cupboard, I I don't mind a little bit of hyper-personalized, helpful marketing. Method.
Until data sets are complete and biases are eliminated, the final guardrail that marketers and their agencies must build and strengthen is the last-mile audit process: (assuming these humans are doing the right thing and don't have the same bias). I'd also like to think that, as exemplified by Dove, we've evolved enough as an industry to know when to limit the types of areas in which we introduce AI into marketing campaigns. We invite the seductive siren voice of AI to lull our industry into the past, to debias everything we do and create, and to do the difficult (and present) (ongoing) efforts should not be undermined.
