Getting Weird With AI

I’ve been getting weird with my AI lately.

Not just a little weird. I’m talking about the kind of weird where you find yourself having a perfectly serious conversation with a 17th-century pirate about internet marketing, arguing with an artificial intelligence about the merits of cats versus dogs, and exploring the physics of eating spaghetti.

And somewhere in the middle of all that nonsense, something interesting happens.

The creative juices start flowing.

Ideas emerge. Conversations take unexpected turns. Connections appear between things that seemingly have nothing to do with one another.

Some of those ideas are ridiculous. Some are surprisingly useful. And occasionally, an idea I hadn’t considered comes along.

But here’s what I’ve discovered.

Getting weird with AI isn’t a guarantee of brilliance. It’s an invitation to creativity.

And I think that’s an invitation more people should accept.

When Did Weird Become a Bad Thing?

Let’s start with the word itself.

Weird. Strange. Unusual. Bizarre. Difficult to explain.

For some people, being called weird is an insult. For others, it’s a badge of honour.

Personally, I’ve always had a soft spot for weird people.

They’re often the most interesting people in the room. They ask questions nobody else thinks to ask. They challenge assumptions. They see connections between things that appear completely unrelated.

They’re the people who look at something everybody else accepts as normal and ask, “Why?”

Or, even better, “What if?”

Think about artists like Andy Warhol, whose unconventional approach helped redefine what art could be.

Or Hunter S. Thompson, who famously wrote:

“When the going gets weird, the weird turn pro.”

I love that line.

And in a world where artificial intelligence is challenging so many established ways of doing things, it seems particularly appropriate.

Because when familiar methods stop working, the ability to experiment with unfamiliar ones can become a genuine professional advantage.

The Problem With Being Normal

It’s easy to approach AI the same way we approach Google.

They type something like:

“Write me an email.”

“Summarize this article.”

“Give me five ideas for marketing my business.”

“Help me write a proposal.”

And the AI dutifully responds.

Usually with something perfectly reasonable, grammatically correct, politely worded, and utterly forgettable.

There’s nothing wrong with those prompts. For straightforward tasks, conventional prompting is often exactly what you need.

I’ve covered those basics in Prompt a Better Prompt. This is what I try when a clear brief still leaves me staring at a predictable answer.

But there’s a problem when conventional prompting becomes the only approach you know.

When you ask an ordinary question in an ordinary way, you’re often inviting an ordinary answer.

And when millions of people use similar prompts to generate marketing materials, business proposals, articles, and social media posts, something predictable happens.

Everything starts sounding suspiciously alike.

The same polished introductions. The same predictable conclusions. The same relentlessly enthusiastic corporate vocabulary.

You know the stuff.

“In today’s rapidly evolving digital landscape…”

Oh, please.

If I see that phrase one more time, I may have to throw my computer out the window.

And I rather like my computer.

The Bathroom Mirror and the Funhouse Mirror

Here’s an analogy that emerged during one of my conversations with my AI.

AI is a mirror.

If you type a generic, normal prompt, you’re likely to get a generic, normal response.

But here’s the funny thing.

Most people are standing in front of a funhouse mirror and asking it to behave like a bathroom mirror.

Think about that.

A bathroom mirror gives you a faithful, predictable reflection. You look into it, straighten your tie, check your hair, and go about your day.

A funhouse mirror is something else entirely.

It stretches you. Distorts you. Makes you look seven feet tall or three feet wide. It shows you possibilities that an ordinary mirror never could.

And sometimes those distortions make you notice something you’ve never noticed before.

That’s what AI can do when you stop insisting that it behave conventionally.

Of course, a funhouse mirror doesn’t necessarily show you something more accurate. Sometimes it just makes your nose look enormous.

But that’s not the point.

The distortion changes your perspective.

And a change in perspective can be the beginning of a new idea.

There’s another dimension to this.

AI doesn’t have to simply reflect your weirdness.

It can amplify it, challenge it, and throw something back at you that you never imagined.

Sometimes the unexpected response itself is valuable.

Other times, it merely triggers a thought in your own mind.

Either way, something has happened that might not have occurred if you’d stuck to the usual script.

I Asked My AI to Become a 17th-Century Buccaneer

Let me give you an example of how this works.

I asked my AI to abandon its usual role as a helpful assistant and become a grizzled, rum-drinking, foul-mouthed pirate captain who had spent forty years plundering Spanish treasure ships.

Then I asked him for marketing advice.

“Arrr, ye want more traffic, do ye, lad?”

“Exactly, Captain.”

“Then tell me, what manner of vessel be this website of yours?”

“It’s a business website. People visit it to learn about my services.”

“And ye just sit there waiting for ’em to come aboard?”

“Well, essentially, yes.”

“BLAST YER BARNACLES! No wonder yer coffers be empty! Ye can’t just float around the Caribbean waiting for a Spanish galleon to sail into yer lap!”

He had a point.

“So what would you suggest?”

“First, ye need a flag!”

“A flag?”

“Aye! A bloody great flag! Something every sailor recognizes from ten leagues away. When they see the Jolly Roger, they know exactly who’s coming!”

“You’re talking about branding.”

“Call it what ye will, landlubber. A ship without a flag is just another bloody boat!”

“Okay. What else?”

“Ye need a reputation! Blackbeard didn’t become famous by handing out business cards at the docks. His name travelled faster than his ship!”

“Word-of-mouth marketing?”

“Aye! Now yer beginning to think like a pirate!”

I was intrigued.

“Give me another idea.”

“Know where the treasure ships sail! Don’t waste yer powder chasing fishing boats when the Spanish treasure fleet be sailing a hundred miles south!”

“Targeting the right audience.”

“Precisely! And when ye find ’em, give ’em a reason to surrender their gold!”

“That sounds a little aggressive.”

“AGGRESSIVE? Ye asked a bloody pirate for marketing advice! What were ye expecting? A complimentary newsletter?”

Now, let’s be clear about something.

Our imaginary pirate hadn’t invented a revolutionary marketing strategy.

Branding, reputation, targeting, and differentiation are well-established marketing principles.

He’d simply presented them through a completely different lens.

But that’s precisely why the exercise was useful.

The pirate made familiar concepts vivid, memorable, and entertaining.

And perhaps more importantly, he got the conversation moving.

What if we took the exercise further?

What if we asked him to apply the actual strategic thinking of a 17th-century buccaneer to modern marketing?

Surprise. Intelligence gathering. Unconventional routes. Strategic alliances. Understanding competitors’ vulnerabilities.

Could some of those historical tactics inspire ethical, modern business strategies?

Possibly.

And that’s where the real creative exploration begins.

There’s a difference between asking AI to sound weird and asking AI to think differently.

The first can make a familiar idea entertaining.

The second may help uncover an unfamiliar one.

Both can be useful, provided we understand what we’re trying to accomplish.

The Real Purpose of Getting Weird

The point isn’t to demand a masterpiece from every prompt. It’s to give my imagination something to react to. Sometimes the AI offers a useful idea. Sometimes its nonsense sends me in a better direction. Either way, the valuable thought may be mine.

Five Ways to Get Weird With Your AI

You don’t need special software, advanced technical knowledge, or a degree in computer science.

You just need a willingness to experiment.

Here are five techniques worth trying.

1. Give Your AI an Unusual Personality

Instead of asking for a conventional explanation, assign an unexpected role.

“Explain why my website isn’t generating leads, but pretend you’re a brutally honest restaurant critic reviewing the worst dining experience of your career.”

Suddenly, your confusing navigation might become an incomprehensible menu, and your weak call to action might become a waiter who never asks for your order.

The analogy might reveal something you’ve overlooked.

Or it might simply make you laugh.

Either outcome can help move the conversation forward.

2. Make Your AI Argue With Itself

Ask for a conventional recommendation.

Then challenge it.

“That sounds exactly like what every one of my competitors would do. Now pretend you’re a rebellious entrepreneur who thinks that entire strategy is obsolete. Tear it apart and propose something radically different.”

You’re not asking the AI to be correct simply because it’s contrarian.

You’re asking it to explore competing perspectives.

Sometimes the disagreement is more useful than the original answer.

3. Combine Things That Don’t Belong Together

Ask your AI to explain customer retention using the principles of a successful rock band.

Or develop a business strategy based on the behaviour of an ant colony.

Or explain a complicated financial concept as though you’re a pirate trying to divide treasure among a particularly dishonest crew.

These strange combinations can produce useful analogies and occasionally reveal unexpected connections.

4. Introduce Bizarre Constraints

“Create a marketing campaign for my business without using the words quality, service, professional, trusted, or affordable.”

Those five words appear in an astonishing amount of small-business advertising.

Removing them forces the conversation in another direction.

Constraints don’t guarantee creativity, but they can disrupt familiar patterns.

5. Ask Questions Backward

Instead of asking your AI to solve a problem, ask it to invent a fascinating problem worth solving.

“Imagine that five years from now, my company has become the most admired business in its industry. What unconventional decision did we make today that everybody thought was crazy at the time?”

Now you’re exploring possibilities instead of merely optimizing what already exists.

And that can be a very different experience.

When the Going Gets Weird, the Weird Turn Pro

Let’s return to Hunter S. Thompson.

“When the going gets weird, the weird turn pro.”

I love that line because it suggests something interesting beneath the humour.

For generations, businesses have rewarded conformity.

Follow established procedures. Respect the hierarchy. Use proven methods. Don’t rock the boat.

And for a long time, those behaviours often produced predictable results.

But what happens when the procedures become obsolete almost overnight?

What happens when artificial intelligence can perform tasks that once required entire departments?

What happens when established business models begin changing faster than the organizations built around them can adapt?

The person who has spent an entire career following the rulebook suddenly discovers that somebody has set fire to the rulebook.

Meanwhile, the weird person is standing there holding a marshmallow on a stick.

There’s something to be said for that.

People who are comfortable experimenting, challenging assumptions, and operating outside conventional boundaries may be better equipped to navigate unfamiliar territory.

They don’t necessarily have all the answers.

And they certainly don’t adapt instantly to every challenge.

But they’re often more willing to explore unfamiliar possibilities.

Of course, businesses ultimately reward results, not eccentricity.

Being weird isn’t an achievement in itself.

But the willingness to experiment can be a valuable asset when familiar approaches stop working.

Maybe the greatest risk in the age of AI isn’t being weird. It’s being relentlessly normal.

Stop Treating Your AI Like the School Principal

There’s another behavioural dimension to all of this that I find fascinating.

Watch how people interact with AI.

Many approach it cautiously, almost deferentially.

They ask polite questions, receive polished answers, and accept those answers as though they’ve just been handed down by an authority figure.

It’s almost like being back in school.

The teacher knows everything. The student raises a hand and hopes not to ask a stupid question.

But that’s a terrible model for creative collaboration.

Your AI isn’t your school principal.

You don’t need permission to challenge it.

You can disagree. Interrupt. Demand alternatives. Introduce ridiculous hypotheticals. Ask it to defend its reasoning.

You can tell it that its last answer was boring.

Believe me, I’ve done that.

You can even tell it to stop being so damned sensible.

And something interesting often happens.

The conversation changes.

Instead of passively accepting information, you begin actively shaping the process.

You become a director rather than merely an audience member.

And your AI becomes something closer to an improvisational collaborator.

That doesn’t mean you should be rude or assume that every unconventional answer is better.

It means you should feel free to experiment.

After all, there are no classmates waiting to laugh at your stupid question.

And your AI is remarkably difficult to embarrass.

The Twentieth Exchange

Here’s something I’ve discovered through working with AI every day.

Some of our best ideas don’t appear in the first response.

Or the second.

Sometimes they emerge after twenty exchanges, when the original question has evolved into something neither of us anticipated.

Recently, my AI and I collaborated on an article called Tickle My Funny Bone, exploring whether artificial intelligence has a sense of humour.

We began with a fairly straightforward question.

Could AI actually be funny?

Not simply repeat jokes it had encountered somewhere, but participate in creating something genuinely amusing?

As the conversation progressed, we began challenging each other, playing with absurd scenarios, and discovering unexpected punchlines.

Some ideas fell flat.

Others made me laugh.

And a few were genuinely surprising.

The most interesting part wasn’t any single joke.

It was the process.

I would introduce an idea. My AI would take it somewhere unexpected. I’d react, change the direction, or add another twist.

And gradually, something new to our conversation began to emerge.

Neither of us started with the finished result.

We discovered it through the interaction.

The magic isn’t necessarily in the first prompt. Sometimes it’s in the twentieth exchange.

That’s a lesson I wish more people understood.

AI prompting isn’t necessarily a one-way transaction where you place an order and wait for the finished product.

It can be a conversation.

An exploration.

An improvisation.

And sometimes, a wonderfully weird adventure.

Nobody Gave Us an Instruction Manual

One reason I believe so many people are still using AI conventionally is that we’re all learning how to use this technology while it’s evolving beneath our feet.

There are tutorials, courses, articles, videos, and increasingly sophisticated prompting techniques.

But there isn’t one universally accepted curriculum for becoming an effective AI collaborator.

We’re still discovering what works.

And much of that discovery comes from experimentation.

People who work with AI every day have an opportunity to develop instincts that occasional users may not yet have acquired.

They learn when to be specific and when to leave room for interpretation.

When to demand precision and when to invite imagination.

When to accept an answer and when to push back.

And perhaps most importantly, they learn that the quality of the interaction depends on much more than the quality of the first question.

That’s why I think we need to broaden the conversation about AI literacy.

It’s not enough to teach people which buttons to push or which prompts to copy.

We should also be teaching curiosity, critical thinking, experimentation, and creative collaboration.

We need to teach people how to play.

Because play is often where discovery begins.

The Human Is Still the Captain

There’s one important distinction that I don’t want to get lost in all this talk about pirates, funhouse mirrors, and creative experimentation.

The human remains the final decision-maker.

Always.

AI can generate ideas at astonishing speed.

It can explore alternative perspectives, challenge assumptions, and produce unexpected combinations.

But it can also produce nonsense.

Sometimes wonderfully entertaining nonsense.

Sometimes dangerously convincing nonsense.

And sometimes an idea that sounds brilliant until you examine it for more than thirty seconds.

That’s why creative exploration needs to be followed by critical evaluation.

Let your AI be weird during brainstorming.

Make it explain its assumptions, check the important facts against reliable sources, and test the idea on a small scale before you spend a dime.

Ask whether the idea is practical. Whether the assumptions are sound. Whether the facts are correct. Whether the proposed solution actually addresses the problem.

And be prepared to discard ideas that don’t survive scrutiny.

The goal isn’t to replace human judgment.

It’s to give human judgment more possibilities to consider.

Weirdness expands the possibilities. Human judgment determines which possibilities deserve to become reality.

Think of the process in three stages.

Explore. Evaluate. Execute.

First, open the door to unusual ideas.

Then examine them critically.

Finally, put the promising ones to work.

Not every idea deserves to survive all three stages.

And that’s perfectly fine.

The Serious Business of Being Weird

I’m not suggesting that you ask your AI to behave like a deranged circus clown when preparing your tax return.

There are times when precision, consistency, and conventional methods are exactly what you need.

If I’m asking AI to analyze financial information, review technical specifications, or help troubleshoot a software problem, I want accuracy.

Not a philosophical interpretation delivered by a tap-dancing pineapple.

But when I’m brainstorming, developing marketing strategies, exploring business opportunities, or trying to communicate an idea in a memorable way?

That’s when I want the funhouse mirror.

I want surprises.

I want uncomfortable questions.

I want unusual comparisons and unexpected connections.

I want my AI to challenge my assumptions instead of simply agreeing with them.

And I want to be willing to challenge its assumptions in return.

Because the real objective isn’t weirdness for its own sake.

Getting weird with AI isn’t about getting ridiculous answers. It’s about discovering extraordinary answers to questions you wouldn’t ordinarily think to ask.

And sometimes, it’s simply about getting unstuck.

If you’ve ever stared at a blank page waiting for inspiration, you know how valuable that can be.

You don’t have to strike gold every time you dig.

Sometimes you just need someone to hand you a shovel.

Your AI Is Waiting for You to Get Weird

So here’s my suggestion.

The next time you open ChatGPT, Claude, Grok, or whichever AI you prefer, resist the temptation to begin with another perfectly ordinary request.

Try something different.

Give it an unusual role.

Ask it to challenge your thinking.

Introduce a ridiculous constraint.

Have it examine your business from the perspective of someone who has never encountered human civilization.

And when it gives you an answer, don’t automatically stop there.

Push a little further.

Ask another question.

Change the rules.

See where the conversation takes you.

You might discover that some of the most valuable insights emerge when you abandon the familiar path.

Or you might simply have a good laugh and return to the original problem with a fresh perspective.

Both are worthwhile.

Because the real secret isn’t that AI suddenly becomes brilliant when you ask it to get weird.

It’s that you give yourself permission to think differently.

And that might be the most important thing of all.

After all, if AI really is a mirror, there’s a rather intriguing question worth considering.

What happens when the person looking into the mirror finally decides to get weird?

I suggest you find out.

And if you happen to encounter a rum-soaked pirate offering marketing advice, tell him I said hello.

Just don’t let him manage your advertising budget.