Generative AI’s Potential & Pitfalls For Advertising Campaigns
Learn how advertisers are leveraging generative AI to create unique campaigns.
Matt McAllister, founder and managing director of Candor Content, shares how brands such as Coca-Cola, Mint Mobile and Alaska Airlines are experimenting with generative AI for recent advertising campaigns. And how the results show the technology’s potential for disruption, but with it comes plenty of pitfalls as well.
Coca-Cola has a long history of collaborating with famous artists, from Norman Rockwell and Andy Warhol to N.C. Wyeth and Gil Elvgren. But for the new artwork on its two most prominent billboards, the brand is calling for help from an unconventional source—inviting fans to design submissions with the help of generative AI.
In a campaign launched recently dubbed, “Create Real Magic,” Coca-Cola has partnered with Bain & Company and OpenAI to build a platform that it claims is the first of its kind allowing artists to create ad campaigns using GPT-4 and DALL-E, the generative AI tools that produce text and images, respectively, from human prompts. Fans from select countries can use the platform to design artwork using Coca-Cola’s branded assets—everything from its famous Spencerian script logo to its iconic polar bears—as a canvas for their AI experimentation. A panel of judges will select the best submissions to be featured on billboards in New York’s Time Square and London’s Piccadilly Circus.
On top of driving fan engagement, the campaign is intended to reaffirm the brand’s position as a vanguard of the cultural zeitgeist by tapping into the recent frenzy around generative AI. It also speaks to the company’s belief that generative AI will transform advertising departments’ operations. According to Global Chief Marketing Officer Manolo Arroyo, Coca-Cola plans to leverage OpenAI’s technology to test everything from content creation and iteration to personalized content and consumer engagement, all in the very near future.
“We’re just scratching the surface of what we believe will help create the industry’s most effective and efficient end-to-end marketing model,” he said in a company blog post. “We will begin to leverage OpenAI’s technology in our marketing function to re-imagine how we produce creative content.” Arroyo expects that doing so will help the team iterate much more rapidly and develop more designs faster than possible through human efforts alone, thereby “increasing the velocity from weeks to days.”
See More: 4 Ways AI Is Changing Content Creation and Marketing as We Know It
Like generative AI itself, however, the “Create Real Magic” campaign is not without risk. For instance, fans might try to create images or text that bastardize the Coca-Cola brand in some way. There are also IP laws to worry about: much of the material that major generative AI tools are trained on comes from sources that could potentially lay claim to the content’s ownership, a fact that has already sparked several lawsuits and is raising questions in Congress about the Copyright Act. Then there’s the overall unpredictability of generative AI. You never know when a chatbot might churn out something insensitive, if not downright racist or sexist. Even ChatGPT’s creators admit that AI can sometimes be “biased, offensive and objectionable.”
To combat these concerns, Coca-Cola worked with OpenAI to build a proprietary version of the latter’s AI tools, resulting in what the brand is calling a “custom-created sandbox” for artists to play in. The platform appears to have been trained only on brand-safe material, and artists must begin by choosing from a curated gallery of the company’s own assets. The platform also seems intelligent enough to ignore requests that could generate restricted content, such as anything that might be construed as offensive or that might run the company afoul of any copyright issues. Precautions such as these aren’t foolproof, but they are a necessity if a major brand like Coca-Cola is going to leverage such unproven technology.
Generative AI As An Ad Copywriter
While Coca-Cola might be the first major brand to invite fans to collaborate on the creation of generative AI content, it is far from the only one experimenting with the technology.
In January, Mint Mobile released an ad for which it supposedly used ChatGPT to write the copy. The video opens with Ryan Reynolds, movie star and part owner of the discount wireless company (which has since been acquired by T-Mobile), explaining that Mint Mobile is always looking for ways to save money. Reynolds then implies that by using ChatGPT, he could save the money that he usually would have spent on a copywriter. Whether this is true or not, it cuts to the heart of the fact that ChatGPT threatens to steal jobs from ad copywriters and other creative types—a potentiality that might be cause for celebration by cost-conscious advertisers but one that raises plenty of questions about the role of creatives in an AI-powered world and whether generative AI is truly up to the task.
In the one-minute video, Reynolds explains that he asked the chatbot to write a commercial for Mint Mobile in his own voice, and to do so using a joke, a curse word, and by letting people know that the company’s holiday promotion is still happening. The copy that ChatGPT came back with was surprisingly good. So good that one commenter questioned whether the copy was actually written by AI or by Ryan himself. So good, too, that Reynolds called it “mildly terrifying.”
Why “terrifying”? Presumably, either because Reynolds finds it scary that artificial intelligence could write something so human-like or that it could write something sounding so much like his own unique voice. After all, Reynolds is one of the thousands, if not millions, of celebrities and content creators whose work can be easily replicated by generative AI solutions. Many of those content creators find it terrifying to see their work replicated by AI, and even more so because someone else is able to capitalize on their work without compensating them in any way.
Regardless, the ad was a resounding success, garnering 1.7 million views on YouTube in its first two months. But the questions it raises about the obsolescence of advertising copywriters, the copyright issues involving AI, not to mention the sheer eeriness of AI’s anthropomorphism, won’t be settled any time soon.
See More: Combatting the Risks of Generative AI
Not All Generative AI Scripts Are Created Equal
Of course, generative AI isn’t always so spot on, and many of the advertisers who have tried it did not get as encouraging of a result. Alaska Airlines, for one, claims to have asked an AI bot to write a commercial for the Super Bowl spot it planned to run with Tan France of the hit show “Queer Eye.” The script turned out to be so banal and uninspiring that the brand shared it on YouTube while dryly commenting, “we have a few notes.” For the actual commercial that ran during the big game, Alaska Airlines went in an entirely different direction, one presumably written and conceived of by real humans.
These early tests of generative AI for advertising campaigns are just the beginning. With the current flood of excitement around the technology and all of the practical applications that it can help with, the global market size for generative AI is expected to top $110 billion by 2030, according to Acumen Research and Consulting. As long as generative AI continues to deliver on its ability to speed up creative iteration, reduce time-to-market, and automate repetitive tasks, advertisers will find ways to incorporate it into their workflows.
While they do so, however, they must be aware of the pitfalls that come along with the technology. Brand safety must be taken into account. Systemic biases must be addressed. Intellectual property and copyright issues must be debated. Agencies and creatives must find ways to adapt to the new reality.
That will all take time. For now, we can tell just from looking at the initial efforts from brands like Coca-Cola, Mint Mobile and Alaska Airlines that the promise of generative AI for advertising is obvious, but so is the peril.
How do you think Generative AI will transform advertising? Please share with us on Facebook, Twitter, and LinkedIn. We’d love to hear from you!
MORE ON GENERATIVE AI
