As the stock photography industry continues to advance through AI integrations with image-generating platforms, such as OpenAI’s DALL-E 2 and LG’s EXAONE, these platforms enable creative teams within ad agencies to unleash their imaginations, producing a library of ad assets in seconds simply by typing in a prompt. This revolutionary capability is available at the cost of the platform’s monthly subscription, making high-quality, customized visual content more accessible than ever before. Big brands are embracing this: “Advertising executives indicate early experiments by the likes of Nestle, Unilever and Mondelez in collaboration with ad giant WPP show generative AI’s potential for slashing the time and money required for marketing campaigns. The technology can rapidly generate text, images and other creative assets that appear original while adhering to brand guidelines.”

Another example is Ogilvy’s campaign for Nestlé’s The Milkmaid yogurt in France, which demonstrates the creative potential. When Vermeer’s famous The Milkmaid painting was in the news after hidden details were revealed via X-ray, Ogilvy used DALL-E 2 to generate AI-imagined extensions of the original canvas. The buzz around the painting allowed the AI-powered video to go viral for the brand without any media spend.

Last week, colorist Laszlo Gaal developed a mock “Volvo – For Life” commercial using the text-to-video RunwayML Gen-3 Alpha AI platform, from concept to final edit, in less than 24 hours. While this project was purely a demonstration of Mr. Gaal’s talent and ability to push the platform’s capabilities, the execution is nearly flawless and could easily air on network TV without compromising the Volvo brand. This groundbreaking technology expands the original concept of stock photography into the realm of video, offering unprecedented possibilities for content creation that is faster, cheaper and more flexible. Just as stock photography revolutionized the way images were sourced and used, text-to-video AI is poised to transform the production of commercials, which can easily have a starting cost of $300,000, making text-to-video AI more accessible, versatile and cost-effective.

However, as the toy retailer brand Toys “R” Us found out, poorly executed AI can also backfire. When it used OpenAI’s Sora to create an entire commercial, it flopped.

The commercial tells the story of Toys “R” Us founder Charles Lazarus and mascot Geoffrey the Giraffe, starting from a humble garage in the 1930s and culminating in a magical portal filled with 3D-rendered toys. Here’s a very harsh review of the commercial: “As a piece of advertising, this thing is an abomination. But it’s not all on Sora—much of the blame rests on the humans behind the AI. Let’s start with the obvious: Toys “R” Us chose to tell a story that is neither interesting nor well-told. No matter how cool Sora’s visuals could have been, the commercial’s conceit did not set it up for success. That’s pure human error. Then there are the visuals, which are both stilted and creepy—not quite real enough to veer into the Uncanny Valley, but strange enough to signal to viewers that something is off.”

These AI platforms are also starting to resonate within the entertainment industry, and as shared here, their impact can be felt at all stages of production. “At its core, the technology takes a script or a series of text-based instructions and transforms them into a visual narrative, complete with characters, backgrounds and even special effects. These AI Hollywood tools are not just limited to creating simple animations; they can produce high-quality, engaging videos that rivl traditional video editing methods.

The process for a video generator involves a series of complex algorithms that analyze the text for context, emotion and intent. Based on this analysis, the AI video generator selects appropriate visuals, soundtracks and transitions to create videos that are not just visually appealing, but also emotionally resonant.”