A kindergarten teacher recently asked me for something that sounded simple: a kids’ map with bunnies, rabbits, and carrots. The idea was adorable and practical—she wanted to write each child’s name and date of birth on a carrot, then use it to celebrate birthdays throughout the year. Also keep it in class as a poster.
My first instinct was to use AI image tools to speed things up. I tried three different platforms, and each time I got close—but not close enough. The images were “fine,” but they didn’t have the exact character or warmth she was looking for. It felt like the tools could generate something that looked like a bunny, but not the bunny that belonged in her classroom.
That experience forced me to admit something: AI has real limitations when the goal is emotional fit, not just visual output. Yes, you can prompt more. Yes, you can get more skilled at prompting. And if you’re a truly creative person, you probably have a clear picture in your head of what you want. But when the output doesn’t match that picture, you’re stuck wondering: is the tool limited, or did I just write the wrong prompt?
In this case, the context mattered a lot. This teacher has done this every year without AI. She would clip bunny and rabbit pictures, arrange them like a collage, print it, and then write the kids’ names. It’s hands-on, a bit messy, and full of personality. Ironically, even with AI at my disposal, trying to recreate that same feeling became an extremely cumbersome effort.
What I learned is that AI’s weaknesses show up fast when an image needs to convey emotion and match a specific environment—like a kindergarten classroom. The images I was getting started to feel monotonous. They were polished but kind of lifeless, and honestly a bit prosaic. Not the kind of playful, inviting look that makes a child smile when they spot their name on the wall.
This is why, as much as we’d like AI to be the solution to everything, I don’t think it can replace creativity. AI has a strong left-brain appeal: it’s built to remix patterns from past data and produce something usable. That’s incredibly powerful for search, coding, enterprise workflows, and anything that benefits from pulling together known information quickly and cleanly.
But creativity—especially the kind that works with mood, context, and human connection—doesn’t always behave like a database query. You can ask for a bunny with carrots and get something technically correct, even fancy. You might even get wild variations, like surreal characters and over-the-top styling. Yet when you need something that feels just right for a group of five-year-olds, “technically correct” isn’t the bar.
So my takeaway is simple: AI can help, but it isn’t a shortcut to taste. It can speed up drafts and give options, but it doesn’t automatically understand the emotional target you’re aiming for. And sometimes, the old-school collage approach—made by someone who knows the kids, the classroom, and the vibe—still wins.
Wharton on AI's impact on creativity
Share this post