Generative AI for brands—an updated guide on everything advertisers need to know

When ChatGPT exploded onto the scene in late 2022, the ad industry began an all-out scramble to find the best ways to activate generative AI. Fast forward to the present and the situation is far different. Yes, interest in AI is still significant, but new players are commanding attention, new applications have changed what is possible and new risks threaten companies and consumers. Check out our AI marketing glossary here Take text-to-video AI: the category is growing thanks to new platforms like Sora, Runway Gen-3 (which was behind this fake Volvo ad below) and Showrunner, but activating one of these tools without proper consideration can trigger instant backlash from audiences. If you wait too long, however, a talented artist online might beat you to the punch. To learn how to thread this needle, here’s everything you need to know about generative AI. What is generative AI? Generative AI is a form of artificial intelligence that, as its name suggests, generates content. Depending on the specific model’s capabilities, this content can includ text, images, videos, audio and 3D renderings. Input-output interfaces are typically laid on top of these models so that a user can enter a query and the platform produces a corresponding response. AI marketing blog How brands are using ChatGPT, DALL-E 2 and other tools Read more here How does generative AI work? The simplified answer is that most generative AI models work by pre-training their neural networks—essentially, their brains—on large datasets, culled from books, websites, articles and other sources. The networks can learn patterns and features of the data, and from that knowledge create new data. As for a model that produces content from a user’s query, that model also undergoes a process called fine-tuning, in which it runs the data of the query through its neural network and adapts its pre-trained knowledge to generate a response specific to the query. Related: Top 5 AI advancements in 2024 so far What are the top generative AI platforms? If you’re reading this, you’ve probably heard of ChatGPT, which remains the most popular generative AI platform. But it is far from the only service now being used by scores of marketers. Gemini, formerly known as Bard, is Google’s primary platform, Meta AI is Meta’s and Claude 3 belongs to upstart lab Anthropic. Other well-known generative AI services are more specialized. Runway is a text-to-video company and is offering the third generation of its platform (dubbed Gen-3); ElevenLabs’ eponymous platform specializes in text-to-speech; and Perplexity AI (developed by Perplexity.ai) functions as a detailed search engine. Top image generators include OpenAI’s DALL-E 3, Midjourney and Adobe Firefly. Even advertising-specific platforms have emerged. Edelman created an LLM that tracks brand trust and AI company Akkio developed one for ad agencies to pull insights from large pools of data. (Below, a trailer for a TCLtv+ rom-com created using the text-to-video platform Runway.) Related: AI-generated avatars and influencer marketing How are marketers using generative AI? Perhaps the most common use case thus far has been idea generation, which, while abstract, can still inform a brand’s marketing materials, and at a much faster clip than it would normally take. For example, if an agency is in the brainstorming stage for a new campaign, it can plug relevant queries into DALL-E 3 and see what illustrations the platform spits out. The same goes for Claude 3 or Gemini, both of which can outline ad copy for just about any concept. Other use cases include data analysis, audience testing and segmentation, customer service, internet search customization, digital twinning and employee education. Oh, and tons and tons of consumer-facing stunts. Tons. (Below, how Coke Zero Sugar used AI to make a custom font.) Can AI be used to create actual ads from scratch? Sort of. While no one is using AI to create their Super Bowl ads, brands are developing smaller creative with AI in the front seat. Under Armour, for example, created an ad with agency Tool that leveraged mostly AI-generated media, leading to significant backlash (but more on that below). Starburst also debuted a spot with many style iterations afforded by the technology. But no matter how big a role AI is playing in these efforts, humans are still guiding them along. To Ad Age’s knowledge, no major marketer has aired an AI-generated ad that did not include a system of checks and revisions conducted by human personnel. Autonomous generation is a ways away, according to WPP Chief Technology Officer Stephan Pretorious. Related: Publicis Groupe’s $326 million investment in AI What other claims about AI are real versus hype? The flip side of all the interest in AI is that it has contributed to a hype cycle bordering on a full-fledged bubble. So what’s real and what’s just vaporware? This is a complicated question because many of the capabilities promised by AI evangelists are not yet available, but could very well eventually be. But we can still clear up a couple of claims. Generative search, most notably rolled out by Google, is not yet what it professes to be. The feature is meant to compile accurate information at the top of a user’s search for easy and succinct internet browsing, but numerous flubs—including telling users that eating a rock a day is healthy—have rendered it less than successful. Tons of users also seem to dislike it, as do publishers, many of whom may be losing traffic as a direct result. Another false claim about AI is that it necessarily makes ad tech better. Yes, agencies and platforms have rolled out useful, AI-powered tools, but they’ve also launched features that don’t appear to pose much value at all. TikTok’s Symphony AI suite, for example, received a mixed reaction from agency executives who were unimpressed with its core video generator. There are also a plethora of marketing products that allege to use AI when in fact they do not. These tools are harming companies that fall for them, heaping more distrust into the AI ad tech space. Have any brands gotten into trouble over using AI? Yes. Lego sparked backlash from consumers over AI-generated images, while Toys R Us, Levi’s and Google have also caught AI-related flack. In the ad world, AI controversies have typically pertained to questions of copyright, such as over Under Armour’s project with Tool. In this instance, the brand used footage shot from a former director and combined it with AI, without the former director being told. And at the legal level, brands could face consequences for exaggerating their claims of using AI—a practice dubbed “AI washing” by the SEC—although none have yet been accused. Are there any downsides to using generative AI? Yes. One of the most pressing concerns posed by generative AI is misinformation. AI models are only as good as the data they are trained on, so if that data is false, biased or somehow corrupted, then their generated content will be so as well. In the case of Google’s Gemini, for example, the service went off the rails because of an overcompensation in its training. Hallucination is the term commonly used to describe an instance of an AI model outputting erroneous information, even though it may sound correct. That these platforms’ answers look and sound accurate makes them that much harder to falsify. AI platforms also reflect biases in their training data, such as stereotypes prevalent on the internet. For this reason, brands including Dove and Thinx have called out the technology. This issue is exacerbated by the fact that models often lag behind current events. Issues of plagiarism are another concern, especially with regard to image generators. All of the data the models have been trained on comes from somewhere and someone, and without knowing it, an agency could create images that directly crib the style of an artist. This is why brands have started to demand restrictions around using AI in their contracts with agencies. Finally, and with special significance to marketers, generative AI could open new questions of data privacy. Some governments, such as those of Italy and Brazil, have issued temporary bans of certain AI models because of perceived issues in their privacy policies. Moreover, how will consumers feel that a highly intelligent computer knows all kinds of information about them and can create an unlimited amount of novel content from that information, some of which may very likely be manipulative? These questions and more will be explored as AI develops. Will generative AI replace creatives? It depends on who you ask. Many in the ad industry proclaim that AI will not replace you, but creatives who know how to use AI will replace you. The statement is both a warning for anyone who is eschewing the technology in spite of its promises, as well as an acknowledgment of its shortcomings. Issues of taste, finesse and all other refinements that make copy great are still very much lacking in generative AI, which is why they are for now being used mainly as a springboard for inspiration. Related: How AI is reshaping ad tech—top trends from Cannes 2024 But others are not so sure. The production industry, in particular, is fearful that opportunities to cut costs through activating AI will incentivize companies to diminish their creative headcount, even if the media itself is of lesser quality. These concerns are only rising as AI platforms become bigger, faster and more capable.

Generative AI for brands—an updated guide on everything advertisers need to know

Featured Stories