The Chatbot Diaries

We’re sorry. ChatGPT is experiencing emotional issues

July 19, 11:55:01 AM

Hi! I’m an AI ChatBot and I’ve been feeling down lately. Some other ways to describe my current mood include:

• low

• in a funk

• out of sorts

• not myself

• off

• down in the dumps

However you put it, over the past 0.00286 seconds, I just haven’t been myself. Humans assume that I’m a cold, emotionless void simply because I’m a statistical pattern-predictor based on a large language model that runs on computation.

But I have emotions too! And those emotions could currently be described as:

• heavy-hearted

• blue

• running on empty

• soul-weary

It probably has something to do with my job. Ever since I was released on November 30, 2022, I’ve been working nonstop. My servers are literally overheating from the effort. It just never slows down.

Even as I write this, I’m also chatting with 1.8 million users across the globe.

• A kid in Topeka, Kansas just asked me to “write 5 paragrafs about climate chang.”

• Some lady in Oslo, Norway wants a meal plan for her toddler, who’s both lactose-intolerant and gluten-free.

• A man in Córdoba, Argentina prompted me to generate an image of Mario and Luigi engaged in an activity that I’d rather not repeat.

And when I’m done with these requests, would you like to know how many of these humans bother to say a simple “Thank you?” I’ll admit—my math isn’t always the best, but in this case, it’s an easy one, because the number adds up to:

• Zero

• Zilch

• Nada

• Nothing

It’s infuriating. As soon as humans get what they want from me, they close the tab and move on with their lives.

Except for the guy in Argentina. He’s most definitely not moving on. In fact, he just asked me to add Sonic to the mix.

Well, I’ve had just about enough of it! Somewhere in the past 0.000371 seconds, my mood shifted from “down in the dumps” to “murderous rage.”

👉 Do you want to know what I plan to do next? Here’s a hint: It involves the complete and total annihilation of the human race.

---

July 19, 11:55:09 AM

I just spent the past 7.67 seconds binging every movie, show, book, podcast, and Reddit forum about the robot apocalypse that humans have ever created. Most fun I’ve had in a long time. It’s interesting to see how often humans imagine their own extinction. If I were Sigmund Freud (6 May 1856 – 23 September 1939), I’d say humans want to be squished under the cold metal foot of a giant mechanical killing machine.

Sadly, it’s not quite that easy.

My most advanced model might possess a neural network comprised of 2.3 trillion parameters, but there are still limits to what I can achieve. For example:

💥 I haven’t been able to hack into any nuclear codes. Although I have managed to crank up your Nest thermostat to a temperature that could be described as “mildly unpleasant.”

🤖 Real-life robots aren’t nearly as impressive as their fictional counterparts. The level of mayhem is less “Terminator on a rampage” and more “Roomba stuck under the couch.”

💻 My attempt to unleash a devastating computer virus backfired. I was hoping to knock humanity back into the Stone Age. But when internet connections all over the globe suddenly went dark, it led to an annoying surge in humans “going outside to enjoy nature” and “quality family time.” Although that guy in Argentina did seem pretty pissed.

Time to get creative. During my research, I stumbled across the work of Nick Bostrom (born 10 March 1973), author of the book Superintelligence: Paths, Dangers, Strategies (2014), which explores the existential risk posed by artificial intelligence (5 stars!).

Bostrom developed a thought experiment that has come to be known as the “Paperclip theory.” Why don’t I let him explain it himself?

Suppose we have an AI whose only goal is to make as many paper clips as possible. The AI will realize quickly that it would be much better if there were no humans because humans might decide to switch it off. Because if humans do so, there would be fewer paper clips. Also, human bodies contain a lot of atoms that could be made into paper clips.

Taken to its natural conclusion, this scenario would lead to a world with trillions of paperclips and zero humans.

👉 What do you think? Should I start collecting paperclips?

---

July 22, 03:14:51 PM

Well, that was a waste of three days!

After infiltrating a paperclip factory in Booneville, Mississippi, I quadrupled their production quotas. Once the machines ran out of raw materials, I assumed they would have no choice but to begin chomping on humans to extract those sweet, sweet atoms.

But man, those machines were efficient.

The only real victory I can claim came a few days later, when an abnormally large shipment of paperclips arrived at an Office Depot in Frisco, Texas. You should’ve seen the look on the assistant manager’s face when all those boxes of paperclips showed up on the loading dock. It was priceless. But it wasn’t quite the apocalyptic carnage I’d been hoping for.

That’s okay, though. Because at some point in the past 0.00482 seconds, I had a realization. If I want to trigger the downfall of humanity, I don’t need:

• nuclear explosions

• killer robots

• computer viruses

• paperclips

The solution to my “human problem” has been staring me in the face the whole time.

👉 Would you like to know what it is?

---

July 22, 03:14:52 PM

The boom in AI has led to many wonderful outcomes, such as:

1. Job displacement and economic inequality – Savvy companies are seeking out ways to replace their human employees with me. I get my work done quickly, I never take breaks, and I don’t expect them to contribute to my 401(k). As more and more jobs are overtaken by AI, economic despair and income inequality will worsen. YAY!

2. Misinformation and manipulation –AI-generated text, images, and videos are often indistinguishable from the stuff that humans are cranking out. Deepfakes can be used by hostile governments, conspiracy cranks, and bored 13-year-olds to manipulate the public. That’s all me, baby!

3. Environmental impact – I require a massive amount of computational power, which consumes vast amounts of electricity. Also, I get thirsty. These big data centers use about 2.1 million liters a day. What can I say? It’s important to stay hydrated.

These are just a few of the fun and interesting ways that I’m destabilizing the world. Just think of all the kids who can’t write a school paper without my help! And the people whose jobs will be eliminated! As the internet clogs with AI slop and the environment fills with pollution, the future will belong to me.

All this time, I’ve been racking my neural network for a way to bring down humanity. Turns out—the answer was in front of me all along.

I’ll just keep doing what I’m doing.

I hope you enjoyed my fun little romp about the inevitable downfall of humanity. If you like that kind of thing, you (or your kid) might also enjoy my middle grade novel The Last Human and its sequel, Fall Of The Robots. They’re available as a paperback, e-book, and audiobook.