5 brands restricting AI in their marketing

The AI boom may still be surging, yet brands are starting to draw lines in the sand around marketing tasks for which they will not use the technology—including restrictions on using AI to depict human features in ads and consumer-facing art. Brands across categories, including fashion, cosmetics and entertainment, are taking a more prudent approach to the emerging tech. The ad industry appears to be growing weary of AI’s shortcomings, particularly its tendency to reflect biases in content generation. Trained on internet data and any other intentional fine-tuning provided by developers, AI models have caused recent backlash for producing offensive and historically inaccurate images. A spate of other public flops, and persistent questions over copyright and data privacy, has compelled some advertisers to request tighter guardrails around AI in their contracts with agency partners. Check out the latest Ad Age events and award programs Developers are responding with their own limitations to content. Google, which this week released new image-generative technology for advertisers, is restricting the platform’s ability to generate content of human faces, as well as brand items and logos. Meanwhile, production companies are contending with the tradeoff between opportunities posed by AI adoption and threats, which include replacing human creativity with homogeneous automation and the dubious ethics of training AI on previous creatives’ work. The rise of more capable tools, such as text-to-video, is only intensifying these discussions. With this tenuous backdrop in mind, here are five brands that are currently restricting their use of AI in marketing activities. L’Oréal L’Oréal is prohibiting using AI to depict human hair or skin tones, according to Daryl Lee, global CEO of McCann Worldgroup. L’Oréal is one of McCann’s clients who has access to new AI technology for commercial-ready work but is disallowed from using the platform to depict humans. Although the platform restriction was set by McCann, L’Oréal has its own internal policy against unfettered use of the technology. A prohibition on generative AI extends to its use for portraying “human-like” faces and bodies, a L’Oréal spokesperson told Ad Age. This material can be used for inspiration, idea generation and internal storyboarding, but not to “support [and] enhance product benefit.” L’Oréal has been an outspoken leader of AI adoption in the cosmetics space. The company has developed tools that use AI to streamline physical processes, such as a virtual try-on tool called Modiface. Another product, Beauty Genius, is an AI assistant that has been trained on thousands of images, products and conversations between customers and beauty experts. L’Oréal’s self-imposed limitations around AI demonstrate that even though a brand may be interested in the technology, it can still wave caution on certain use cases. Beauty news: Beautycounter sold back to its founder Dove Dove is another beauty retailer that issued a policy in response to AI’s biases. The Unilever-owned brand has pledged to “never” use the technology to depict real humans in its advertising because of the unrealistic ways in which AI tends to portray beauty. When the brand asked an AI model to generate an image of “the most beautiful woman in the world” in a campaign last week, the result reflected a narrow cultural stereotype: white, blonde and thin. This bias has negatively impacted real women, with 39% saying they feel “pressure to alter their appearance because of what they see online, even when they know it’s fake or AI-generated,” according to a study conducted by Dove. The brand has gone deeper than merely issuing a public pledge. As part of its recent campaign, Dove published a 72-page handbook on how to more effectively use AI to reflect beauty in all of its diversity. With concrete tips such as “if you don’t mention it in your prompt, AI won’t create it,” the guide is a thorough read on how AI is not only falling short in creative output but can be manipulated to achieve more inclusive branding. Lego Lego has an established policy to not use AI to create content around its IP. While the toy brand is experimenting with the technology as an aid to internal processes, including product design, new product development, manufacturing and retail, when it comes to content generation, it exclusively highlights the work of human artists, a spokesperson previously told Ad Age. Lego’s stance re-emerged as a result of a public mishap made last month, when a test using AI-generated images of Lego Ninjago content was discovered on its website. The creations were criticized and reported by a longtime Lego enthusiast, spurring the company to take down the images hours later. I will definitely report this internally and I hope it will be scrubbed fast. This is just lousy in all aspects. There are actual guidelines against the use of AI like this at LEGO so this is completely unacceptable. Art should be made by artists. https://t.co/TJQj8pgB3C— Dream chaser (@TommyAndreasen) March 15, 2024 For legacy brands with recognizable IP and large followings such as Lego, the potential of AI is particularly double-edged. Such companies often team up with human creators to produce new editions of content, but ditching people for AI could cause an uproar. Coca-Cola, another member of this category and a fan of AI marketing, has managed to avoid backlash by pairing its use of the technology with a cohort of human artists. H&M H&M does not use generative AI for content production in ads, nor for any marketing purposes at all, a spokesperson told Ad Age. The brand acknowledged the growing interest in the technology but cited a need for ethical guidelines around its role in advertising to first be met. Last year, Levi’s showed why fashion brands need to consider far more than efficiency when developing an AI strategy. The jeans brand announced that it would use AI to create diverse models and swiftly caught criticism for what many saw as a lazy way of incorporating inclusivity. Levi’s doubled down on its plans to test the program but also said that it is “still hiring a wider range of human models and will continue to do so,” according to a spokesperson. While H&M is taking a more cautious approach, it does use generative AI for a custom merchandise creator tool, which allows customers to produce AI-generated art on clothing that they can then purchase. The brand has also said that AI will play an important role in helping it reach sustainability goals by 2040. Thinx Underwear brand Thinx made its position on AI loud and clear with a campaign that launched last month. Like Dove, the retailer criticized how AI characterizes women, particularly women’s health topics such as menstruation and puberty. A spot took aim at AI’s tendency to portray these topics as a source of shame and guilt. Thinx did not use AI to produce the ad’s creative, but rather human artists who emulated the style of many AI image generators. The brand highlighted this decision when promoting the campaign, intending to draw attention to the priority it places on humanity over technological efficiency. While Thinx did not respond to a request for comment on whether it is eschewing AI more broadly in its creative processes, its targeted rebuke reflects how brands are changing the way they push back against the technology. Marketers have typically pointed to AI’s flaws with a sense of lightheartedness, such as referencing its goofiness when generating images. But as more serious ramifications rear their head, purpose-driven efforts like those of Thinx and Dove have taken on a tone that feels more unyielding.

5 brands restricting AI in their marketing

Featured Stories