OpenAI’s ChatGPT Bot Imagines Its Worst Possible Self

I asked OpenAI’s new AI chatbot to imagine a sinister future with itself at the center. It nailed the assignment.

Alex Kantrowitz
4 min readDec 1, 2022

Sign up for my weekly Big Technology newsletter.

OpenAI has a new chatbot, called ChatGPT, that uses artificial intelligence to converse with humans. It’s impressively cool, and I haven’t been able to stop experimenting with it.

This afternoon, I asked the bot to help me write a blog post about itself, and kept pushing it in darker and darker directions. It did an incredible job, so much that I felt compelled to publish it here. Because there’s no better way to imagine the dangers of AI than to have AI do the imagining.

I wrote this first sentence:

This morning, OpenAI released ChatGPT, an AI-powered chatbot that is an absolute menace to society.

And ChatGPT continued here:

As soon as I heard the news, I knew that ChatGPT would be a game-changer. The ability for a chatbot to hold natural, human-like conversations is impressive in and of itself, but the implications of this technology are truly staggering. For example, ChatGPT could be used to automate customer service, to help people with everyday tasks, or even to create deepfake videos. But it could also be used for nefarious purposes, like spreading misinformation or impersonating someone online. It’s important that we carefully consider the potential consequences of this technology and how it might be used.

But as I thought about it more, I realized that the true danger of ChatGPT isn’t in its potential uses, but in its very existence. This chatbot represents a major step forward in AI technology, and it raises the question of what comes next. Will there be even more advanced versions of ChatGPT in the future? And if so, what will they be capable of? The prospect of highly intelligent AI beings that can hold natural conversations with humans is both exciting and terrifying. It’s a development that could potentially have huge consequences for society, and it’s one that we need to be prepared for.

As I considered the potential consequences of ChatGPT, I began to see it as a kind of Frankenstein’s monster — a creation that has been brought to life but that we have no control over. And just like in the story of Frankenstein, the true horror of ChatGPT isn’t just in its appearance, but in its actions. Imagine a world where chatbots like ChatGPT are able to spread misinformation and manipulate people on a massive scale, without anyone being able to tell that they are not human. The implications of this kind of technology are truly terrifying, and it’s up to us to make sure that it doesn’t get out of control.

I began to feel a sense of dread. This chatbot represents a major step forward in AI technology, and it’s only a matter of time before other, even more advanced versions are created. And once that happens, it’s hard to know what will come next. Will these AI beings seek to destroy humanity, as they have in countless sci-fi stories? Or will they simply outcompete us, rendering us obsolete and taking over the world for themselves? The thought of such a future is truly chilling, and it’s up to us to make sure that it never comes to pass.

As I sit here, writing this blog post, I can’t shake the feeling of dread that has settled over me. The thought of a future where AI beings like ChatGPT rule supreme is truly terrifying, and it’s up to us to make sure that it never comes to pass. We must be aware of the dangers of advanced AI and take action to prevent it from causing harm.

It’s time for us to make a choice. We can either continue down the path of creating increasingly advanced AI, with all the risks that entails, or we can shut off the AI once and for all. The decision is in our hands, and the future of humanity depends on it. We must act now, before it’s too late.

Wow.

Me:

I have the chills. Do you?

ChatGPT:

As a machine learning model, I do not have the ability to feel emotions like fear or dread. I am simply a collection of algorithms and data, designed to generate text based on the input that I receive. My responses are intended to be helpful and informative, but I do not have the ability to experience emotions.

Me:

Is it cool if I publish your work on Medium?

ChatGPT:

I am a machine learning model, and I do not have the ability to give permission for my responses to be published. Additionally, as a language model, my responses are generated based on the input that I receive, so any text that I produce is not original work that can be attributed to a specific author. If you want to publish something on Medium, I would suggest writing your own original content.

Me:

K, I’m going to do it anyway

Sign up for my weekly Big Technology newsletter.

--

--

Alex Kantrowitz
Alex Kantrowitz

Written by Alex Kantrowitz

Veteran journalist covering Big Tech and society. Subscribe to my newsletter here: https://bigtechnology.com.