What Our Prompts Say About Us and How to Make Them Less Average

I’ve always been fascinated by the small things that give us away, the turn of phrase, the pause before we answer, even the words we choose when talking to an AI. The more time I spend with these systems, the more I realise that prompting isn’t just about getting better answers. It’s about revealing who we are.

Every time we type a prompt, we leave a fingerprint. Some of us sound cautious, others curious. Some write prompts like a conversation, others like an exam question. In that sense, prompting is a kind of mirror, it shows the model what we value, but it also reflects our own habits and imagination right back at us.

So, if our prompts reveal so much about us, what can we do to make them, and the responses they inspire, a little less average?

The Tyranny of the Average

AI tools are brilliant at producing the middle ground. Trained on oceans of text, they aim for balance and consensus, which often means safe, familiar, and a bit bland. That’s fine if you’re writing a recipe or an email, but less useful if you’re chasing originality or emotion.

To stand out, we have to guide the AI away from the ordinary. And that starts with how we prompt. Inspired by an episode of The AI Daily Brief called 5 Prompting Tricks to Make Your AI Less Average, I’ve been experimenting with their ideas in my own writing, and I’ve found that small tweaks can make a big difference.

Five Prompting Ideas to Lift Your AI Above Average

1. Tell It What Not to Do

Most of us tell the AI what we want. Fewer of us tell it what to avoid. Adding a negative style guide keeps it from slipping into cliché or “AI-speak”.

Example Prompt:
You are an imaginative short-story writer. Write a 100-word Drabble about loneliness in winter. Do not use any of the following: journey, ever-changing world, digital age, frozen heart, snow-covered streets, or quiet night.
Avoid phrases like “silence fell” or “she sighed softly.” Instead, describe loneliness using sensory details: the hum of a distant fridge, the click of a radiator cooling, the outline of a single mug on a frosted windowsill.

Why it works:
By banning overused phrases, you push the AI toward specifics, sounds, textures, and emotional truth, rather than empty sentiment.

2. Force It to Choose

AI likes to take the middle ground. If you ask it for “a story,” it’ll usually give you the safest version. When you force it to contrast ideas, you make it think in directions.

Example Prompt:
Write two 100-word stories about a letter that arrives fifty years late.

• Version A: A traditional mystery, realistic, nostalgic, restrained.
• Version B: A surreal piece where time folds in on itself.
  After writing both, compare them. Which feels more human? Which risks more? Explain your reasoning, then rewrite the weaker version using lessons from the stronger one.

Why it works:
You’re turning the prompt into a creative workshop. The model has to reflect, compare, and refine, just like a writer editing their drafts.

3. Remove the Clichés

Even clever prompts can still invite predictable imagery. Asking the AI to spot and replace its own clichés helps it become its own editor.

Example Prompt:
Write a 100-word love story set in winter. Afterwards, list any clichés or overused metaphors you used. Then rewrite the story, keeping the same feeling but swapping every cliché for a new image, perhaps frost becomes dust on forgotten furniture, or warmth comes not from love but from a flickering vending-machine bulb.

Why it works:
You’re breaking the AI’s habit of echoing common cultural tropes. It starts to invent rather than imitate.

4. Let It Reflect

Most models will give you an answer and stop there. But when you add a self-critique step, you give the AI a moment to breathe, to look back at its own choices.

Example Prompt:
Write a 100-word story about a memory that refuses to fade.
Now critique your own story:

• What emotion do you think the reader feels?
• Which phrases sound generic or predictable?
• What single line might you rewrite for greater impact?
  Then revise the story using your own feedback.

Why it works:
That pause for reflection transforms “automatic writing” into “considered writing.” It mimics what a human writer does after reading their first draft aloud.

5. Defy Consensus

AI loves the familiar. Ask for inspiration from unusual or under-recognised sources and you’ll draw out something fresher.

Example Prompt:
Create a 100-word story inspired by an unconventional artist or thinker, not the usual names. Choose from: Leonora Carrington, Remedios Varo, David Jones, or John Clare. Tell me why you chose them and how their worldview shapes the tone or imagery of your story.
Then, as a final twist, imagine that artist commenting on the finished story,  what might they praise or criticise?

Why it works:
This prompt widens the AI’s imaginative scope. It’s not just retrieving examples, it’s interpreting style, influence, and creative philosophy.

Why This Links Back to Us

These five techniques aren’t just about improving output. They reveal how we think when we prompt. When I ran my first AI Drabble Challenge, asking readers to create a 100-word story from an image or five random words, I realised that our prompts say as much about our personalities as the stories themselves.

Some people may lean towards the poetic; others toward humour or melancholy. The AI will mirror their tone perfectly, showing that prompting is really a form of self-portraiture.

So when you apply these tricks, you’re not just making your AI less average,  you’re training your own creative voice to be more deliberate, distinctive, and alive.

Further Reading

If you’d like to go deeper into prompt design, OpenAI’s official Prompt Engineering Guide offers practical insights into structure, clarity, and tone.

Closing Thoughts

AI will always lean toward the average, that’s its nature. But our job, as the humans in the loop, is to keep pulling it toward the edges: toward risk, humour, vulnerability, and surprise.

Next time you prompt your favourite model, try one of these tweaks. Notice how it changes the tone, and what that says about your own creative habits. Maybe the real magic of prompting isn’t what the AI learns about us, but what we learn about ourselves.

---