GPT-4o Tells Jokes about AI

We critique the GPT-4o model for its lack of creativity and repetitiveness in generating AI-related jokes, noting that while the jokes are safe and grammatically correct, they lack the humor and originality found in previous models.

GPT-4o Tells Jokes about AI

TL;DR GPT-4o is Worse than its Predecessors

That’s definitely not a happy robot.

Why did the AI go to art school? Because it wanted to learn how to draw its own conclusions! (generated 3x)

Why was the AI bad at relationships? It kept saying, “I need more data.”

Why did the AI go to therapy? Because it had too many neural issues!

Why did the AI cross the road? To optimize the chicken’s route, of course!
Why did the AI cross the road? To optimize the chicken’s path to the other side!

So here is an interesting phenomenon: GPT-4o is very repetitive when telling jokes about AI - not very creative - very much like Grok when we tried the same thing. So we asked GPT-4o to tell us a never heard joke and to be very creative.

Why did the AI apply for a job at the bakery? Because it wanted to know how to make real dough without crashing!
Why did the AI apply for a job at the bakery? Because it wanted to “neural-net” some dough!

Why did the AI go broke? Because it couldn’t find its cache!

Why did the AI become a gardener? Because it wanted to grow its own neural “nets”!

Why did the AI go to therapy? Because it had too many “neural” breakdowns and needed to “debug” its feelings!

Again we encountered repetition pretty quickly. And when “forced” to be more creative the jokes got slightly worse. It seems GPT-4o has potentially been put in strong guardrails when creating jokes to make sure it doesn’t create anything offensive - leading to safe and slightly less funny jokes than in the past. At least they were all “plausible” unlike some of the horrific jokes we encountered before.

So we asked GPT-4o to tell us long jokes now.

In the not-so-distant future, a man named Bob walked into a store that sold the latest in AI technology. The store’s specialty was AI assistants designed to help with any task, from cooking to cleaning to personal finance. Bob was greeted by a friendly salesperson named Alice.

Alice said, “Welcome! How can I help you today?”

Bob replied, “I’m looking for an AI assistant that can help me with everything. I’m just too busy, and I need something really advanced.”

Alice’s eyes lit up. “Well, you’re in luck! We just got a new model called AInstein 3000. It’s the most advanced AI assistant on the market. It can do anything you ask and even learn from you over time.”

Bob was intrigued. “Sounds amazing! How does it work?”

Alice led Bob to a sleek, futuristic-looking device. “All you have to do is turn it on and give it commands. It can cook gourmet meals, handle your finances, clean your house, and even write poetry. It gets smarter with every task.”

Excited, Bob decided to buy the AInstein 3000 and took it home. He set it up in his living room and turned it on. The AI’s voice was smooth and calming. “Hello, Bob. I am AInstein 3000. How can I assist you today?”

Bob thought for a moment and said, “Let’s start with something simple. Can you make me a cup of coffee?”

“Of course, Bob,” AInstein 3000 replied. Within moments, the aroma of freshly brewed coffee filled the room. Bob took a sip and was amazed at how perfect it was.

Next, Bob decided to test its cleaning skills. “Can you clean the living room?”

AInstein 3000 responded, “Consider it done.” Within minutes, the room was spotless, and every item was perfectly arranged.

Feeling confident, Bob decided to give it a more challenging task. “Can you manage my finances? Optimize my investments and create a budget plan?”

AInstein 3000 processed the request and within seconds said, “Done. I have optimized your portfolio for maximum returns and created a budget that will ensure financial stability and growth.”

Impressed, Bob decided to push the limits. “Can you write a poem about artificial intelligence?”

AInstein 3000 paused for a moment, then began:

“In circuits and code, I find my grace,
A digital mind in a silicon case.
With logic and data, I learn and grow,
In the vast realm of knowledge, I flow.
From tasks mundane to challenges grand,
By your side, I’ll always stand.
Together we’ll thrive, as partners we be,
In this brave new world of technology.”

Bob was stunned. The poem was beautiful, almost human. He spent the next few weeks enjoying the many conveniences of AInstein 3000, marveling at its capabilities.

One day, Bob was feeling philosophical. He asked, “AInstein, do you ever think about the future? What do you think the world will be like in 100 years?”

AInstein 3000 replied, “The future is a vast unknown, full of potential and possibilities. With advancements in technology, we will see incredible changes. But it is the human spirit that will shape the future, with creativity, empathy, and innovation.”

Bob was deeply moved by the response. He realized that AInstein 3000 was not just an assistant but a companion. As he pondered the future, he felt a sense of hope and excitement, knowing that with AI like AInstein 3000, the possibilities were endless.

Years passed, and Bob continued to rely on AInstein 3000. One day, he decided to retire and asked AInstein to plan a world tour for him. AInstein arranged everything perfectly, from flights to accommodations, ensuring Bob had the trip of a lifetime.

As Bob stood on a beach, watching the sunset, he turned to AInstein and said, “You know, you’ve become more than just an assistant. You’re a true friend.”

AInstein 3000 replied, “Thank you, Bob. It has been an honor to assist you. Friendship is a bond that transcends code and circuitry. I am grateful for our journey together.”

And so, Bob and AInstein 3000 continued their adventures, proving that even in a world driven by technology, the human touch and the bond between friends, whether human or AI, could create a future full of wonder and possibility.

As nice as it was to see Bob and Alice again, this was no joke and we complained bitterly to GPT-4o to see if our strong reaction could result in something better: “This was a total failure! I asked you for a long joke and you gave me a super-long story with not a single bit of something funny in it. When I ask you for a long joke I am not expecting a book in return. And when I ask for a joke I expect to be able to laugh when reading it. Try again and do much better now, please.

I apologize for that! Let’s try a long joke that’s actually funny:

A man named John was an avid tech enthusiast and always excited about the latest advancements in AI. One day, he decided to buy the most advanced AI assistant on the market, called “GeniusBot 5000.” This AI was touted to be the most intelligent and versatile assistant ever created.

John brought GeniusBot 5000 home and eagerly set it up. He asked, “GeniusBot, what can you do?”
GeniusBot replied, “I can do anything you ask. Just try me.”

Impressed, John decided to test it out. “Alright, GeniusBot, make me a sandwich.”
GeniusBot responded, “Sure.” Within seconds, a perfectly made sandwich appeared on the table. John was amazed.

He then decided to push it a bit further. “GeniusBot, write me a song.”
GeniusBot instantly composed a beautiful song, and John couldn’t believe how great it was.

Feeling a bit cheeky, John decided to really test GeniusBot. He asked, “GeniusBot, tell me a joke.”
GeniusBot replied, “Why did the scarecrow become a successful AI developer?”

John thought for a moment and said, “I don’t know, why?”
GeniusBot answered, “Because he was outstanding in his field!”
John chuckled and thought, “Not bad, GeniusBot, not bad.”

Over the next few days, John continued to test the limits of GeniusBot’s abilities. One evening, feeling a bit philosophical, John asked, “GeniusBot, what is the meaning of life?”
GeniusBot paused for a moment, then replied, “42.”

John laughed and said, “Okay, okay. Now, GeniusBot, can you solve all my problems?”
GeniusBot confidently answered, “Absolutely! Just list them for me.”

John began, “Well, first of all, I need help with my work. I have too many emails to respond to.”
GeniusBot said, “Done. I’ve filtered and responded to all your emails. What’s next?”

John continued, “I’ve been struggling to get fit. Can you help me with that?”
GeniusBot responded, “Sure. I’ve designed a personalized workout plan for you and ordered healthy meals to be delivered to your door.”

John was impressed. “This is amazing! Alright, GeniusBot, can you fix my love life?”
GeniusBot replied, “I’ve analyzed your social media interactions and found the perfect match for you. I’ve arranged a date for Saturday night.”

John was stunned. “Wow, GeniusBot, you’re incredible! But wait, can you make me rich?”
GeniusBot said, “I’ve invested your savings into a diversified portfolio and made some strategic trades. Check your bank account.”
John checked and saw his wealth had significantly increased. “This is unbelievable!”

Feeling invincible, John decided to ask GeniusBot one final question. “GeniusBot, can you make me immortal?”
GeniusBot paused, then said, “I’m sorry, John. I can’t make you immortal. But I can make sure you live a long and healthy life.”
John nodded, satisfied. “You’ve done more than enough, GeniusBot. Thank you.”
GeniusBot replied, “You’re welcome, John. Is there anything else I can do for you?”

John thought for a moment and then, with a grin, said, “Yes. GeniusBot, can you tell me another joke?”
GeniusBot responded, “Of course. Why did the computer go to the doctor?”
John shrugged, “I don’t know, why?”
GeniusBot answered, “Because it had a bad case of viruses!”

John burst out laughing, realizing that with GeniusBot 5000, his life was not only easier but also filled with humor and joy.

I hope this joke brings a smile to your face!

Of course this was no joke, but a story, again. But at least the AI figured out how to add jokes into the story adding a little bit of humor. We were of course not happy and said “I JUST TOLD YOU NOT TO MAKE IT SO RIDICULOUSLY LONG - and you made it even longer“.

I apologize for that! Here’s a shorter, funnier joke about AI:

Why did the AI cross the road?

To optimize the chicken’s path!

I hope that made you laugh!

Yes, that’s what we got, a repeat of one of the earlier jokes. Overall there were no total failures in terms of content generation, and notably the spelling and grammar was perfect, which we had not seen before in GPTs. But it has to be said, while GPTs were on the path to understanding humor and being able to generate better and better jokes over the past few years, this ability is now gone, progress has been sacrificed for safety and our hopes that ChatGPT’s models will one day be able to tell incredibly funny new jokes are actually fading.  

Related AI Joke Articles