🧠 Issue #23 – When AI Tries to Be Funny: What Laughter Teaches Us About Being Human

"I'm hungry," my five-year-old announced before dinner last week.

"Hi Hungry, I'm Dad," I replied without missing a beat.

The words escaped before I could stop them—this joke, of all the possible dad jokes, slipping out in a moment of pure parental autopilot. I braced for the inevitable groan.

Instead, he paused, processing. Then came that slow, delighted grin as the wordplay clicked. The kind of smile that says he's just discovered something wonderful and slightly mischievous.

It hit me: this terrible, wonderful dad joke works not because it's clever—it's the opposite of clever—but because it's ours. A ritual of affection disguised as wordplay. A way of saying "I'm here, I'm silly, and I love you" without the vulnerability of actually saying it.

Which got me thinking: we've taught machines to beat chess grandmasters, predict weather patterns, and occasionally write decent poetry. Now they're trying to make us chuckle. The results? Surprisingly effective—and deeply unsettling. Because while AI can mimic the mechanics of humor, it's never experienced the peculiar joy of dad jokes—never felt that mix of pride and embarrassment when your kid discovers the magic of terrible puns. What happens when something that's never felt affection tries to create the language we use to express it?

🔍 The Spark

Recent studies from USC found that nearly 70% of participants rated ChatGPT-generated jokes as funnier than those crafted by humans. Twitter is ablaze with screenshots of AI quips that land better than your Uncle Bob's Thanksgiving zingers. Professional comedians are quietly experimenting with Large Language Models as creative partners, though many won't admit it publicly.

But here's the twist worthy of a punchline: while AI can generate jokes, research suggests it fundamentally lacks understanding of what makes them funny. In tests matching jokes to cartoons, humans achieved 94% accuracy while AI models managed only 62%. It's like watching someone perform a perfect piano concerto while being completely tone-deaf.

The uncomfortable truth? We're not just witnessing AI learn to tell jokes—we're discovering that humor might be less mystically "human" than we'd hoped.

💡The Insight

Humor isn't just clever wordplay or perfectly timed surprises. It's emotional archaeology—digging through shared experiences to find those universal moments of "oh god, yes, that's exactly it."

• A joke works when it reflects something painfully true

• It connects when it catches us being human together

• And it really lands when it gives us permission to laugh at ourselves

Neuroscientist Sophie Scott's research shows we're 30 times more likely to laugh when we're with others than alone. Laughter is social glue, not just entertainment. Yet studies reveal people consistently rate jokes as less funny when they believe AI created them—even when the content is identical.

The gap isn't just technological; it's existential. AI can pattern-match humor's structure but not its soul. It hasn't lived through awkward first dates, workplace disasters, or the peculiar horror of assembling IKEA furniture. It can simulate the form of relatability without ever having been relatable.

🤯 The Tension

Here's where it gets deliciously complicated: if AI can make us laugh without understanding why, what does that say about laughter itself? Are we just pattern-recognition machines giggling at linguistic formulas?

The research suggests "the subjective experience of humor may not be required for the production of good humor—merely knowing the patterns that make up comedy may suffice". Translation: feelings might be optional for funny.

This is either liberating or terrifying. If humor is just elaborate pattern-matching, then maybe authenticity is overrated. Or maybe we're about to flatten one of humanity's most complex emotional experiences into algorithmic entertainment, losing something vital in the translation.

The real punchline? We might be laughing ourselves out of what makes laughter meaningful in the first place.

🔗 Curiosity Clicks

1. Philosophy of Humor – Stanford Encyclopedia of Philosophy Three millennia of philosophical wrestling with what makes us laugh—and why most thinkers have been surprisingly grumpy about it.

2. AI Humor Generation Research Recent breakthrough study showing how giving AI "humor skills" helped it create jokes that nearly matched human-written comedy.

3. Comedy Writers Experiment with AI Real professional comedians share their experiences using LLMs as creative tools—from skepticism to cautious collaboration.

💬 Quote That Hits

🧭 The Human Prompt

Next time something makes you genuinely laugh—not just exhale through your nose, but actually laugh—pause and dissect the moment:

What made that funny? Was it the surprise? The recognition? The permission to acknowledge something we all feel but rarely say?

Notice how much of humor depends on timing, context, shared understanding. On being seen. Because the difference between AI humor and human humor isn't just technical—it's the difference between mimicking connection and actually connecting.What just happened that made this funny? Was it clever? Relatable? Was I surprised… or seen?

Because chances are, that moment said something real about your fears, your friendships, your worldview. Humor isn’t just entertainment—it’s emotional intelligence in motion.

🤔 Worth Considering

We may someday build AI that can riff with us in real time. But it will never feel that awkward silence after a joke flops. It will never experience the thrill of a room erupting in laughter—or the relief of laughing through tears.

And maybe… that’s the point.

The essence of being human isn’t how fast we can respond. It’s how much we can feel when we do.

Let’s keep our laughter real. Let’s make the future more human.
— Jesse