When AI Makes Everything Look Good, Creative Work Wins On Tension and Truth

Before generative AI, visual perfection was expensive, requiring time, talent, and specialized craft. Today, polished imagery can be produced in seconds. As execution becomes faster and cheaper, aesthetic quality alone no longer differentiates. That shift raises new risks for marketers leaning too heavily on automation, where speed can come at the expense of meaning. In response, creative value is moving toward what machines struggle to replicate: the human fingerprint, defined by intentional imperfection, emotional tension, and lived experience.

Guan Hin Tay, APAC Regional Director for The One Club for Creativity, has spent decades navigating industry shifts at agencies like JWT and BBDO. He brings a clear perspective on what makes creative work resonate, shaped by both practice and theory. As the best-selling author of Collide: Embracing Conflict to Boost Creativity, Tay has long argued that tension is a key driver of engagement. As the industry adapts to automated content generation, he sees human flaws and emotional intelligence as the remaining strategic edge.

"Generative AI gives us a perfection that is now a commodity. What used to be slow, handcrafted, and special has now been cheapened by speed," says Tay. This dynamic is showing up most clearly in how creative decisions are made. Cost pressure at the executive level is pushing teams toward faster, higher-volume output, often at the expense of distinctiveness. In some cases, brands are leaning into AI to produce more content, but without a clear creative standard, that volume fails to resonate. The result is inefficiency rather than scale, where speed alone does not translate into impact. "Some clients just want it fast," Tay notes. "You might produce ten pieces using AI, but if nothing actually reaches the consumer, the client will eventually become frustrated."

• Synthetic silent night: The gap between polish and feeling became clear in recent holiday advertising. Coca-Cola experimented with AI-generated Christmas campaigns, but audience reactions pointed to a lack of emotional authenticity. While the visuals met a high technical standard, they struggled to connect. For Tay, the takeaway is straightforward: AI can replicate aesthetics, but not the lived experience that makes nostalgia resonate. "Think about those Coca-Cola Christmas ads that were heavily driven by AI. People still hate them because they feel fake," he observes. "People want to see the Christmas that lives in all our memories, which is something really special. When you have that genuine warmth coming through, that's what people really want from those ads, not just some gimmicky, nice visuals."

• Curating the crooked: Navigating this shift requires creators to rethink where they add value. The advantage is no longer in using the tools, but in directing them. AI can enhance strong creative instincts, but only when paired with clear judgment and editorial control. Often, the difference shows up in subtle choices that move the work away from perfection and toward something more human. "If you input a prompt to create a face, you could maybe make the nose a bit crooked, maybe the eyes are not perfect on the left or the right," Tay advises. "The human element in terms of directing and including our taste, judgment, and editorial skills is critical."

As the line between synthetic and real content continues to blur, editorial judgment becomes essential. Tay notes that some AI-generated images are now convincing enough to pass as real, even to trained eyes. He views AI as a useful layer for brainstorming and research, but emphasizes that human oversight is what preserves the integrity of the final work. While he would use AI for research today, he points out that fully automated content quickly triggered backlash and platform restrictions. "When ChatGPT was launched, you could see people publishing tons of books because it was still new," he observes. "But as we progress and learn more about AI, we can spot these nuances."

• Tension wins attention: That pushback reinforces a broader idea at the center of Collide. Tay’s framework is built on the belief that tension drives engagement, especially as creative work becomes easier to produce. As generative tools flatten execution, what stands out is not polish, but friction, those moments of disruption that make a story feel real. In that context, vulnerability becomes a core ingredient, while overly refined outputs often struggle to connect. "In creative writing, we're always looking for the tension behind the story," he says. "Finding the disruptive tension allows us to build interesting, surprising stories that people want to be engaged in over and over again."

Viewing AI through a storytelling lens also shapes how Tay thinks about its trajectory. Much of today’s development is focused on making language and image models behave more like humans, but imitation is not the same as experience. This gap becomes especially clear when it comes to emotion, where AI can replicate signals without fully understanding them. Tay frames that limitation through a familiar sci-fi perspective. "AI is great at mimicking emotions. It does try and behave as human as possible. If you listen or if you watch any science fiction movie, the advanced AI always tries to be human," he says. "It's quite interesting that humans are creating AI to be human. But then to be human, you also need to be flawed."

Taking that perspective into account brings the focus back to what brands can control today. While the long-term trajectory of AI continues to evolve, the immediate priority is preserving the human elements that give creative work its distinctiveness. That means protecting the imperfections, nuance, and emotional texture that machines struggle to replicate. "As long as AI doesn't think for itself or create its own personality, then hopefully, we are quite safe," he concludes.