AI Humor? I Asked ChatGPT to Write Something Funny & Here is What I Got

My ChatGPT experiment turned into a deep dive into AI-generated comedy and what it means for creators.

An AI walks into a bar...

Have you heard that joke? Neither have I, because AI can't actually walk. But they can try to tell jokes about it!

Come to think of it, an AI-powered android could walk into a bar and tell a joke, so there’s that.

One of my favorite ways to dig into things is to put my experiments where my mouth is. unfortunately, I don’t have access to an AI-powered android, but that would be an awesome way to test things. Poor me, I don’t even have a robot vacuum thingy.

But on the subject of humor and artificial intelligence, stick around and I’ll share with you how ChatGPT responded when I gave it the task of writing something funny. It actually surprised me. But first…

---

🖥️ A Word About AI

When I was a kid, my mother told me I should become a lawyer because I could always argue more than one side. To her dismay, probably, I never became a lawyer.

Instead, I’m a technologist with an optimistic side, a pragmatic side, and a healthy dose (I hope) of realism thrown in. Confession: I have a love-hate relationship with technology, including AI. And of course, I can see both sides.

A few months back, I was reading a lot of (not so funny) stories about how people were afraid of AI, hated AI for stealing people’s work—lots of doomsday-ish stuff. Not surprising since AI is infiltrating our lives and threatening jobs and livelihoods in a way other technologies haven’t in many years.

Once upon a time, Alexa and Siri seemed cool. All we had to worry about were the potential privacy issues (you did worry about that, right?).

The cool factor wears off in a hurry, though, when your career is in jeopardy because someone wants to replace you with a machine. It’s all fun and games until junior grows up and threatens to burn your house down.

There are issues with AI from different angles, no doubt. It will revolutionize many fields and change the landscape of society as much as or more than, computers did. Some of the problems will be sorted out in the courts, other’s through regulation, and the rest we’ll have to figure out as artificial intelligence progresses and evolves into whatever it becomes.

Who knows, maybe AI will become smart enough to solve the problems it causes, or maybe it will go off the rails and ruin everything.

As I was noticing all the AI-focused negativity, though, it hit me. In the tech world, we just automatically reach for the next tool. Like grabbing a butter-laden piece of toast because it’s calling to you (I’ve heard toast does that to some people).

To a tech, AI is the next step in the evolution of tools. It’s a better hammer, if that hammer had the power to potentially put all of humanity at risk in ways no one completely understands yet.

Regardless of any possible downside, it’s expected that as techs, we’ll use it. Like slipping on a new pair of Nike running shoes, we just do it (just kidding, Nikes never fit my feet right).

In fact, in many work environments, you’d better be using AI or you’re not doing your job. It’s expected by your boss, or if not, your boss’s boss.

On a positive note, I find AI super handy for all manner of tasks and projects. ChatGPT is more than just a writing assistant. Among other things, I use it for:

• Coding

• Goal setting

• Trip planning

• Keyword research

• Competitor analysis

• Writing short stories

• Overcoming writer’s block

• Task and project planning and prioritization

• Brainstorming short story ideas and characters

• Editing, summarizing, and condensing documents

• Creating topic clusters for Substack SEO (search engine optimization)

I often bounce ideas off of it and cut down my research time. It’s great if you have to read long documents or watch videos and just want the highlights.

I built a custom GPT as a strategic advisor. It can even give you suggestions on how to butter up a coworker when delivering bad news.

Speaking of coworkers, AI is changing how people work in a big way. I’ve seen firsthand how it can improve business communications and eliminate third-party services that overcharge and underdeliver. With that out of the way, I’m going to focus on the topic of the day, which is humor.

---

😃 Introduction to AI Humor

AI is a hot topic and discussions about it can get heated quickly, with strong views on both sides. To be different, let’s dive deep into whether machines can genuinely make us laugh. And if so, what does that mean for the future of both artificial intelligence and human creativity?

Big question, right? As a writer, author, and creator myself, I understand the mix of fascination and trepidation panic many of us feel about AI's growing capabilities in, and infiltration into, creative fields.

It's one thing for AI to crunch numbers, help us code, summarize documents, or play chess, but humor? That's supposed to be a uniquely human trait, right?

Well, not exclusively human, as some animals pull pranks just for fun. But we’re talking about machines here. How do 1s and 0s transform themselves into a joke?

---

🤔 Why is That Funny, Anyway?

Here’s a fundamental question: What makes something funny? Humor often relies on:

• Unexpected connections.

• Cultural context.

• Timing.

• An understanding of human nature.

Now add in humor being subjective. Can AI really grasp these nuances?

I mean, it's just a prediction engine, right? Can it really tell a joke? Or is it just manipulating patterns and using probabilities without true understanding? That leads you to wonder if understanding is even needed at all?

---

✍️ The Evolution of AI Content Creation

AI content creation has come a long way from simple word association games and playing chess. Today's AI systems can analyze vast amounts of human-generated humor, identify patterns, and attempt to recreate them in new ways. And yet, at the end of the day, it’s still just piecing things together, one word at a time. Or is it?

There have been statements made by AI developers that indicate that sometimes, they’re not sure exactly why an AI does what it does. Is it possible that at this point in the evolution of AI, something is going on under the hood that is not understood?

---

😆 Laughing at the Joke vs. Laughing at the AI

But here's where it gets interesting: Sometimes AI jokes are funny because they're not quite right. There's something amusing about an AI trying to understand human humor and missing the mark slightly—like watching a child tell their first joke, or Midge in The Marvelous Mrs. Maisel (great show) learn the art of perfecting a “tight 10” comedy set.

When an AI generates humor that makes us laugh, are we laughing at the joke itself, or at the fact that it came from an AI? It's a distinction that matters more than you might think.

Software developer and indie hacker, Egor Dubchenko, after trying to build an app based on AI jokes, concluded:

I know now that AI is bad at humor, and this is what makes me laugh when I use it.

The rise of AI humor has effectively created a new genre of comedy—one where the line between intentional and unintentional humor becomes blurry. Sometimes, the funniest AI-generated content comes from its attempts to understand human concepts. And not quite getting it.

But is AI really bad at humor?

---

🏫 Survey Said: Machines Are Funnier

On the other hand, there are those who consider AI-generated humor to be on par with or better than what humans create, at least in some cases. A USC psychology study shows that AI-generated jokes are funnier than those created by people. Say what? Yeah, go figure.

“Since ChatGPT can’t feel emotions itself but it tells novel jokes better than the average human, these studies provide evidence that you don’t need to feel the emotions of appreciating a good joke to tell a really good one yourself,” Gorenz said.

One study compared the funniness of novel jokes generated by regular people with the funniness of novel jokes generated by ChatGPT. Participants rated the funniness of all jokes without knowing their author. The researchers found that overall, nearly 70% of the participants rated the ChatGPT-generated jokes as funnier than those crafted by humans. A little over 25% of participants rated the human-crafted responses as funnier, and only about 5% rated jokes from both sources as equally funny. The results were consistent even when the researchers checked for differences across various demographics.

In that same article, a second study was mentioned with a group of 200 participants who rated AI-generated headlines equally funny to those of The Onion. Their conclusion is:

“But for professional comedy writers, our results suggest that LLMs can pose a serious employment threat,” Schwarz noted.

Whether AI gets the last laugh remains to be seen.

That last part indicates some uncertainty about the future, but keep in mind that is only 2 studies.

---

📉 Implications for Creators

For writers and creators, this raises some uncomfortable questions. Namely: “What does my future look like?” If AI can make us laugh, what else can it do? Are our creative jobs at risk? Should we be worried about being replaced by algorithms that can generate endless streams of content?

The truth is more nuanced than simple replacement. AI might be able to generate jokes, but it can't understand why they're funny on an emotional level—at least, not yet. It's pattern recognition versus genuine understanding.

Many AI jokes are surprisingly clever, while others fall completely flat. The interesting part, though, is that both successful and unsuccessful attempts can tell us something about how AI processes information.

---

🧰 Mechanics of AI-Generated Jokes

When AI creates content, it chooses the words it uses based on probabilities. When AI tries to be funny, it's essentially performing a complex mathematical operation based on patterns it's learned from human humor. But isn't that, in a way, what our brains do, too? We learn what's funny through exposure and pattern recognition.

This brings us to another interesting question: If an AI can consistently make humans laugh, does that indicate some form of intelligence we don't fully understand yet?

Put another way, if an AI told another AI a joke, would it think it was funny? How would it determine that? Could it “Rate how funny this joke is from 1-10?” How would it be calibrated?

📃 Examples

To add some flavor and context, I asked ChatGPT to generate some writer-oriented jokes. Here are a few examples:

Setup: “I tried to give my short story a shocking twist.”
Punchline: “The real shock was nobody read it.”

“Writer’s Block is like telling your story in Morse code but losing the cheat sheet mid-sentence.”

Setup: “I took a creative writing class and learned one golden rule: ‘Show, don’t tell.’”
Punchline: “But I’m a writer, so now I just talk about showing instead of actually showing.”

“Every time I feel proud of a paragraph, I realize I borrowed all the words from the dictionary.”

Setup: “Whenever I see a blank page, I feel the weight of infinite possibilities.”
Punchline: “So I quickly add a coffee stain to narrow them down.”

“There’s nothing more humbling than reading your old work. It’s like meeting your past self and immediately wanting to apologize.”

Setup: “Somebody told me writing was a ‘lonely profession.’”
Punchline: “They clearly haven’t met all the characters living in my head.”

Some of those are amusing and clever, others not so much. Those are just one-liners or at most, a couple of lines, though. I wanted to take it further, and that leads us to my experiment…

---

🧪 My AI Experiment

I took a break from whatever geeky thing I was doing and I asked ChatGPT to write something funny for me. Why? To see for myself if what people said was true.

I had heard people say:

• AI is flat.

• AI can’t write anything that would require a human to be a little creative.

The fundamental question: Can AI write something funny from the seed of an idea? If it can, does that indicate some deeper level of understanding? Understanding of what? Human emotion? Human reactions? Or is it just solving a math problem and piecing together a string of words?

To see if AI could really capture humor based on a concept, I gave it a prompt and let it do its thing.

The prompt wasn’t very sophisticated, but was good enough to give the AI the flavor of what I wanted. In my prompt, I told it to, “Write a light hearted, amusing, and at times laugh out loud hilarious recipe card for a cooking website.”*

I went on to say, “Pretend the chef is a French-born, arrogant, and over-the-top personality with amazing cooking skills to match. He speaks English with a French accent and all the attitude of a top-tier famous person in his field.”

I almost said to make him a French Gordon Ramsay, but I stopped myself because I wanted to see what it would do on its own.

I told the AI that the chef likes to insult people and is critical of other chefs. Butter is his favorite thing to add to any dish, and lots of it. For the recipe, I chose a decidedly non-French Texas bar-b-que and some side dishes, and let the AI go to work on it.

And now, the dish you’ve been waiting for, the pièce de résistance…

---

📃 The Results from the AI

Below is the full text of what ChatGPT spit out as a first draft. This is what the AI gave me, plain and simple.

I think it’s scary good. It’s a better first effort than a lot of human writing I’ve seen. Take a look and tell me if you think an AI wrote it.

If it’s this good now, what does the future hold?👇

Chef Jean-Marc Le Magnifique’s “Texas BBQ Ribs à la Française”

Serves: 4 pathetic souls who need guidance in the kitchen.

Prep Time: As long as it takes for you to stop crying about how much butter I use.

Cooking Time: Long enough for you to question why you ever tried to cook ribs without me.

Ingredients for the Texas BBQ Ribs (how tragic, but I will fix them):

• 2 racks of mediocre Texas-style pork ribs (I suppose this will do)

• 2 sticks unsalted butter (non-negotiable, imbécile – don't even ask)

• 1 tablespoon Dijon mustard (the only thing here with dignity)

• 2 tablespoons smoked paprika (Americans think this is fancy—how cute)

• 1 teaspoon garlic powder (fine, if you must)

• 1 teaspoon onion powder (also bland, but we will make do)

• Salt and pepper (use liberally—because at least it isn’t ketchup)

• 1 cup of BBQ sauce (store-bought? Sacrebleu! At least make it tolerable)

For the Texas Toast (please...):

• 4 thick slices of white bread (oh, mon dieu! It’s like eating clouds of disappointment)

• 1 stick of butter (obviously)

• 2 cloves garlic, finely minced (finally, a decent ingredient)

For the Potatoes and Onions (it’s sad I have to include this):

• 4 large russet potatoes (basic, but acceptable if drowned in butter)

• 1 large yellow onion (I will not weep for you, but you might)

• 1 stick of butter (you see the trend, yes?)

• Salt and pepper (as if anything could redeem these vegetables)

Instructions

Step 1: Prepare the Ribs... or at least try.

1. Preheat your oven to 300°F (or 150°C for people who know what they’re doing).

2. Pat the ribs dry with paper towels. Yes, dry—don't let them drip like your enthusiasm for life.

3. In a bowl, mix the Dijon mustard (an ingredient worth my time), paprika, garlic powder, onion powder, salt, and pepper. Slather this mix generously on both sides of the ribs. Generously, I said! You Americans and your fear of seasoning—it’s pathetic!

4. Wrap the ribs tightly in aluminum foil (like a protective layer for your fragile ego) and place them in the oven. Let them roast for 2.5 hours, until they are so tender you might shed a tear. But not from joy—from realizing how much better it could have been if you were French.

Step 2: Texas Toast... I’ll make it edible, don’t worry.

1. While the ribs cook, heat a large skillet over medium heat. Throw in that whole stick of butter—do not skimp! If you think about using margarine, I will find you.

2. Once the butter is melted (ah, sweet nectar), add the garlic and sauté until fragrant. Toast the bread slices in the buttery garlic mixture until they are golden, crispy, and almost worthy of being called food.

3. Remove the toast from the pan and place on a plate. Look at them. Try not to weep.

Step 3: Potatoes and Onions... I suppose.

1. Peel the potatoes, though honestly, the thought of even peeling these is exhausting. Slice them into 1/4 inch rounds. Slice the onion into thin rings.

2. In a large skillet, melt (you guessed it) another stick of butter. So much butter, so little time for your complaints.

3. Add the potatoes and onions. Season them with salt and pepper, and sauté until golden and crispy on the edges. They should be soft enough to make you realize the futility of life.

Step 4: Finish the Ribs – Let’s Make Them Pretend to be French.

1. After 2.5 hours of slow roasting, remove the ribs from the oven. Preheat your grill or broiler to high. Brush the ribs with BBQ sauce (sigh, if we must).

2. Grill or broil the ribs for 5 minutes, just to caramelize that sauce. Turn occasionally, but don’t burn them like you burned your last relationship.

3. When done, brush with—what else?—melted butter. Serve them up with your Texas toast and buttery potatoes and onions, and then quietly reflect on how much better they are now, thanks to me.

Final Note from Chef Jean-Marc:

These "Texas BBQ ribs" are now only vaguely Texan. Thanks to me, they have become a magnificent fusion of the American bourgeois barbecue and French culinary brilliance. Butter, the only ingredient that can save your cooking, has elevated this dish beyond your wildest, cholesterol-filled dreams.

Bon appétit! Or as you say in Texas… Enjoy, y’all. Pfft.

---

📔Philosophical and Creative Considerations

If humor requires intelligence, and AI can create and eventually understand humor, then what does that tell us about the nature of intelligence itself?

Looking at the problem through this lens might actually be good news for creatives and other professionals. The fact that we're even discussing whether AI can truly understand humor shows how far away we are from machines fully replicating human creativity.

If you believe the best comedy often comes from emotional understanding and shared human experiences in a cultural context, then AI can’t replace people. Not yet. AI can mimic these elements, but can it truly understand them? No.

Is a robot stand-up comic in our future?

And yet, there’s that study I mentioned that concludes that understanding isn’t necessary to produce humor. I have to wonder if we’ll have a future filled with AI-created sitcom TV shows and romantic comedy movies. What about a robot stand-up comic? It seems laughable (pardon the pun) or low budget sci-fi now, but it’s amazing what you can get used to.

The current state of AI humor is like watching a gifted translator work with a language they've never actually spoken—they can get the words and the syntax right, but might miss subtle cultural nuances. And subtle differences can have big impacts.

This limitation becomes apparent when AI attempts to create topical humor or jokes that require current cultural context. The results can be anywhere from brilliantly absurd to completely ridiculous and nonsensical.

But here's where it gets really interesting: Sometimes AI creates humor that's funny in ways humans wouldn't think of. Imagine a comedian from another dimension—they see things differently, and that itself can be entertaining.

The fear of AI taking over creative jobs isn't new, but you’d think the peculiarities of humor would actually make it one of the last frontiers. Making someone laugh consistently requires more than just stringing words together in a clever way. Humor also goes far beyond just telling one-liners or writing funny headlines.

---

⌚ Limitations and Challenges of AI Humor

Timing might be AI’s Achilles heel. Some of the best comedians have an amazing sense of comedic timing. AI can be programmed to understand the technical aspects of timing, but can it feel the rhythm of a joke like Carol Burnett or Steve Martin? Not likely.

Then there's the matter of audience awareness. Human comedians can read a room and adjust their material accordingly. AI, for now, can’t.

---

🤝 AI as Creative Partner

An important point about AI humor: As with most things AI, it's currently more of a collaborative effort between human and machine than a purely artificial creation.

The most successful AI content often comes from a back-and-forth with the tool. Human input (naturally) starts the process. Then there is refinement, curation, and selection. That leads to a more refined prompt, and whoever prompts the best, wins.

We all now have at our disposal an infinite number of monkeys with typewriters, but with better spelling and grammar. You just keep making revisions to the prompt until things come out the way you want.

Think of AI as a brainstorming partner with a very unusual perspective. It might not always get things right, but it can help you think outside the box. In my experience, people tend to limit themselves during brainstorming. AI can help remove barriers and take your content in a totally different direction.

---

🔮 The Future: Human-Machine Collaboration

The evolution of AI humor also tells us something about machine learning. As AI systems get better at understanding context and nuance, their attempts at humor become more sophisticated.

But does more sophisticated necessarily mean funnier? Not always. Sometimes the most entertaining AI-generated content comes from its more primitive attempts at understanding human concepts. Sometimes the old AI model does a better job than the current one.

This raises another interesting question: As AI gets better at generating humor, will we lose that "AI trying to be human" quality that makes some of its current attempts amusing? Ultimately, it’s up to people to decide how to best achieve the outcome.

---

🤝 AI as Para-Technology

For writers and creators, the key might be finding ways to collaborate with AI rather than compete with it. After all, who better to write jokes about AI than a human-AI comedy team?

At its best, AI is a para-technology tool, meaning it works alongside the writer, author, or creator, not as a replacement. The future of AI humor might not be about replacing human comedians, but about creating entirely new forms of comedy at speed that wouldn't be possible without artificial intelligence.

Consider the potential for AI to help identify patterns in successful comedy, or to generate unexpected combinations that human writers might not think of. As we get more and more into our intelligent machine era, could this human-machine partnership, perhaps unexpectedly, lead to meaningful comedy?

---

📼 AI and Human Nature

Source: Giphy

Is it live, or is it Memorex?

Does the ability to generate humor imply consciousness? That's a philosophical rabbit hole that could keep us debating for days.

What we can say is that AI humor shows us something about both artificial intelligence and human nature. Our reaction to AI-generated jokes and to AI in general tells us as much about ourselves as it does about artificial intelligence.

It’s not just that we feel AI poses a threat. Part of the reason people fear AI is it can be a brutally honest and cold mirror. Write your best work and then ask AI to pick it apart. It can be a humbling experience. It’s in times like this we begin to question ourselves.

Creators worried about the future can remember this: AI might be able to generate content, but it can't replace the shared human experience. There’s a reason things become memes and go viral: connection.

The best AI jokes often work because they provide a unique perspective on human behavior and society, even if unintentionally.

This unintentional commentary might actually be one of AI's most valuable contributions to humor. It's like having an alien anthropologist studying human culture and reporting back with refreshingly naive observations.

For writers and creators, this presents an opportunity. AI humor can be a tool for exploring new creative territories and perspectives that result in a deeper human connection.

The future of AI humor might not be about machines replacing human comedians, but about finding new ways to make us laugh by highlighting the absurdities in both human behavior and machines.

---

⏭️ What’s More? What’s Next?

As AI continues to evolve, we might need to expand our understanding of what constitutes humor. Perhaps AI will develop its own unique branch of comedy that humans find funny for entirely different reasons.

For now, AI humor serves as a fascinating but sometimes distorted mirror, reflecting our own assumptions, fears, and hopes about artificial intelligence and human creativity.

To creators worried about the future: AI humor is currently more of a tool than a replacement. It's like having a slightly confused but enthusiastic intern who sometimes comes up with brilliant ideas by accident.

The real value of AI humor might not be in the jokes themselves, but in what they can teach us about creativity, intelligence, and ironically, what it means to be human.

Humor, after all, requires emotion in order to get the joke. To truly get it in your core, it requires you to feel something. Machines can’t do that.

As we continue to explore and develop AI capabilities as a whole, humor will likely remain one of the most interesting areas of study. After all, what could be more human than laughing at a robot trying to tell a joke?

So the next time you see an AI-generated joke, think about this: You're watching a piece of history as it unfolds. You're not just reading a piece of content created by algorithms— you're witnessing the ongoing evolution of artificial intelligence and its attempt to get a grip on one of humanity's most complex and interesting traits: our sense of humor.

---

🤔 What Do You Think?

I’m curious to hear your opinion. The good, the bad, and the ugly.

---

*Credit where credit is due, I got the idea for a humorous recipe card prompt from reading some of the material in the Solo Build It! (affiliate link) training. I host one of my websites there and have access to the tools and training. This includes deep dive reports on things like search engine optimization and AI prompting.