Most AI comparisons focus on performance.
Which model reasons better?
Which one writes better?
Which one codes faster?
I wanted to look at something else:
imagination.
What happens when four leading AI systems receive exactly the same open-ended creative prompt?
There is no correct answer.
There is no target style.
And there is very little guidance.
The participants are:
ChatGPT
Microsoft Copilot
Google Gemini
Claude
For each experiment, the four models receive the same basic task.
Sometimes they are asked to imagine a place.
Sometimes they are given a role.
Sometimes they must invent something.
And sometimes they must turn their idea into a single image.
THE RULES
The rules remain consistent throughout the project:
1. The same basic prompt goes to all four models.
2. The question must allow genuine creative freedom.
3. Additional instructions are kept to a minimum.
4. Each model works independently.
5. Results are preserved as they were produced.
6. Unexpected details and mistakes are part of the record.
I am intentionally not introducing the individual personalities of the four models yet.
The results should do that for themselves.
This project is not intended to determine the world’s “best AI.”
Instead, I want to document how different generative AI systems interpret the same creative idea.
It is also a time capsule.
These experiments are being conducted in 2026.
AI is changing so quickly that repeating exactly the same experiments in the future may produce completely different results.
That makes today’s outputs worth preserving.
Perhaps in a few years, we will look back at them and immediately recognize what AI creation looked like in 2026.
So I am turning the experiment into a series.
The prompts will be revealed.
The creations will be shown.
The differences will become obvious.
And eventually, each of the four participants may reveal far more about itself than I originally expected.
For now, however, the laboratory door has just opened.
Four AIs.
One prompt.
Let’s see what happens.
