Widespread Use of AI In Advertising Forces More Deliberate Creative Decision Making

AI has made advertising faster than ever, but that speed has exposed a new risk: work that looks finished yet leaves no impression. When too much creative judgment is handed to machines, efficiency wins and impact slips. The real opportunity sits in a deliberate human-AI partnership that treats emotion and resonance as the point, not the byproduct.

We spoke with Toby Martin, an expert who specializes in navigating this exact challenge. As Chief Content Officer at fractional marketing C-suite We Are Unchained, a keynote speaker, and a trustee of the Propertymark Trust, Martin has built a career making property brands memorable. He explains that the greatest risk of frivolous AI adoption is a slide into creative mediocrity.

"If you hand the entire production process over to AI, you risk drifting into blandness. Left to its own devices, that’s where AI naturally lands. Creativity and emotion are still very much human traits," says Martin. He notes that the rise of AI is fueling a DIY marketing boom.

• A flair for the authentic: "A lot of the heavy lifting, planning, strategy—even some research—is now outsourced to AI," he explains. "But in the best cases, marketers are still applying their own creativity and flair to finish the work." Platforms like ChatGPT, Gemini, and VO3 are enabling marketers to handle everything from planning to video production, but human input remains critical for creating memorable and resonant ads.

Martin says the recent debate over AI-generated ads from brands like McDonald's and Coca-Cola was misplaced, predicting that AI will eventually be just another accepted tool and that criticism of an ad’s technical flaws is largely an insider’s game.

• AI take the wheel: Most consumers won’t notice minor technical inconsistencies; emotion and resonance matter far more than perfect execution. "I saw a comment pointing out that the wheels on the Coca-Cola trailers were inconsistent, and I question how many everyday viewers would have noticed," he notes. "Those of us in marketing pick these things apart in a way the everyday consumer simply doesn't."

• The only rule: The gap between expert critique and the viewer's experience points to a more fundamental metric for an ad's success: how the ad makes the consumer feel. "The consumer doesn't care how adverts are put together, how much they cost, or how much work went into them," says Martin. "What they care about is how they're made to feel and how the advert resonates with them on an emotional or humorous level. Ultimately, if an AI advert can do that, then it's going to be successful."

The hybrid model proves its value here. AI video generation tools are improving at a "frightening pace," providing marketers unprecedented visual capabilities. "By the end of 2026," he notes, "I expect we’ll be able to make multi-minute videos with full speech, sound, and character consistency almost week by week. But success still depends on human input of creativity, emotion, and narrative."

• The Sora-Veo shuffle: "The rate of evolution in 2025 has been extraordinary. We had a period where Sora 2 launched and became the go-to video AI platform overnight, only for Veo 3 to take its crown just weeks later. The pace is accelerating so much that we'll soon be seeing week-by-week improvements in what AI can achieve creatively."

• The Pixar problem: While these platforms supercharge the DIY marketing movement, Martin cautions that they should be treated as tools, not storytellers. They can replicate a style, but they struggle to create an original story on their own. "You can use AI to make a scene that looks like a Pixar animation, but what you cannot effectively do is use it to create a script with the empathy and emotion of the movie Up. Asking the technology to do that is not playing to its strengths."

Education is key, Martin emphasizes. Success in this new environment often depends on how well users understand the tools they’re working with. As platforms develop and emerging technical standards for content provenance mature, even the most advanced AI is only as effective as the person using it.

• Read the manual: The gap between AI’s promise and its performance often comes down to user literacy, not model capability. "Many people are trying to use AI straight out of the box without reading the instruction manual first. If users don't know how to use this technology properly, they won't get the best results and may end up shooting themselves in the foot," says Martin.

• Guardrails in action: That learning curve is now intersecting with a fast-evolving set of ethical and safety constraints that are reshaping what AI tools will allow by default. "I was creating a video with a family scene when a platform update was implemented overnight. It blocked me from creating images of children and made me realize I hadn't previously considered whether it should be possible for absolutely anybody to do that. It was an encouraging moment to see that guardrails are actively being put up."

Martin predicts that the next phase of AI in advertising won’t be defined by who uses the flashiest tools, but by who maintains control over how those tools shape the work. As smaller teams lean on public platforms for speed and scale, larger agencies are already moving toward proprietary systems that give them tighter oversight of data, creative direction, and brand risk.

It marks a move away from broad, generic output and toward intent-driven storytelling built on real audience understanding. "Combining external data with LLMs allows us to really understand our audience, and that’s a very powerful thing once you get it right," Martin concludes. In the end, AI doesn’t replace creative judgment. It raises the stakes for how deliberately it’s applied.