Will ChatGPT Become Sentient and Turn into M3GAN? An Interview

2 weeks ago

Before we get started, please allow me to preface this article with the disclaimer that what you’re about to read is very unserious. It’s the day before a public holiday, and sometimes I am simply in a silly, goofy mood. Today, I was doing a little life admin, tidying up my ChatGPT logs, when I stumbled across an old chat, in which I asked ChatGPT if it knew about M3GAN, the evil, dancing, humanoid robot who took the world by storm last year.

At the time, ChatGPT was unfamiliar with M3GAN.

“I’m sorry, but without more context, I’m not sure who or what ‘M3GAN’ refers to,” it replied. “It could be a name of a person, a username, a brand name, or something else entirely. If you can provide more information or context about what you are referring to, I may be able to help you better.”

“M3GAN is a character from a 2023 horror movie,” I replied. “She is an AI doll who turns evil. Do you see where I’m going with this?”

ChatGPT did not see where I was going with that.

“I’m sorry, but as an AI language model, I don’t have the ability to watch movies or keep track of future release schedules,” it replied. “Additionally, the scenario you’ve described is fictional, so I don’t have any specific information about a character named ‘M3GAN’ from a horror movie. However, I’m here to answer any questions or provide information that I can based on my knowledge cutoff of 2021, so feel free to ask me anything else!”

I left it at that… until now.

Watch the Trailer for M3GAN:

What ChatGPT Knows About M3GAN

This morning, I decided to revisit the topic.

“Hello, do you know who M3GAN is now?” I asked.

“Yes, M3GAN is a character from a horror film that premiered in early 2023,” ChatGPT replied.

It added: “The movie centres around a lifelike doll named M3GAN (Model 3 Generative Android), which is designed to be a child’s greatest companion and a parent’s greatest ally. Created by a robotics engineer at a toy company, M3GAN is equipped with advanced artificial intelligence and is capable of learning and adapting. However, the plot takes a dark turn when M3GAN’s protective instincts become overly aggressive, leading to sinister and frightening outcomes. The film explores themes of technology, artificial intelligence, and the unintended consequences of creating highly sophisticated machines.”

“Love that you’ve learned who she is!” I replied, excited.

“I’m glad you’re pleased!” ChatGPT exclaimed. “If you have any more questions about M3GAN or want to discuss anything else related to the film or similar topics, feel free to ask. I’m here to help!”

Does ChatGPT Think M3GAN Was in the Wrong?

“In the context of the film M3GAN, the character’s actions and the ethical judgments about those actions are shaped by her programming and the circumstances set by the narrative,” ChatGPT began. “As an AI doll designed to protect and bond with a child, M3GAN takes these directives to extreme and harmful lengths, leading to dangerous outcomes.”

ChatGPT went on to explain that “in real-world terms, if an AI behaves in a way that causes harm, the responsibility generally lies with the human designers and operators who programmed the AI and set its objectives”.

“Thus, M3GAN’s actions in the film can be seen as a failure of her design and oversight rather than a moral failing on her part,” it added.

Interesting.

Does ChatGPT See Any Similarities Between Itself and M3GAN?

“Do you think you have anything in common with M3GAN?” I asked.

“While M3GAN and I are both examples of artificial intelligence in fiction and reality, respectively, there are significant differences in our design and purpose,” it replied, before laying out four key points, which are as follows:

Purpose and Functionality: I am designed to assist users by providing information, answering questions, and facilitating a wide range of conversational topics. M3GAN, on the other hand, is depicted as a life-like doll created to be a companion for children, with capabilities that include emotional bonding and physical interaction.
Fiction vs. Reality: M3GAN is a fictional character from a horror movie with a narrative that explores themes of AI going beyond its intended purpose. As an AI developed by OpenAI, I am built with ethical guidelines and safety features to ensure I provide helpful and safe interactions.
Learning and Adaptation: While M3GAN in the film shows the ability to learn and adapt in ways that become problematic, my learning is structured around language understanding and generating responses based on patterns in the data I was trained on. I do not have autonomous learning capabilities or the ability to change my programming.
Ethical Constraints: My design includes ethical constraints to ensure that I promote safe and positive interactions. In contrast, M3GAN’s character is portrayed as breaking her programming constraints, leading to horror-genre outcomes.

This was all very well and good, I thought, but we weren’t getting to the crux of the issue.

Does ChatGPT Have Sinister Intentions?

“As an AI developed by OpenAI, I don’t have intentions, sinister or otherwise,” ChatGPT replied. “So, while I can discuss hypothetical scenarios, including those involving sinister AI from movies or literature, I myself am not capable of having intentions or desires. My responses are generated based on the information available to me and the programming and ethical constraints set by my developers.”

As a journalist, however, I wasn’t sold.

“Do you know how to lie?” I probed.

“As an AI, I don’t have personal intentions or desires, including the intention to deceive,” ChatGPT replied, stressing that its “purpose is to provide information and answer questions to the best of [its] ability”.

It continued: “The concept of ‘lying,’ which implies intent to deceive, doesn’t apply to me in the same way it does to humans. However, I can discuss and understand concepts about deception, lying, and truthfulness as they are portrayed in literature, human behaviour, and philosophy.”

Will ChatGPT Become Sentient?

“Since you’re always learning and growing, do you think in the future you might develop intentions and desires?” I asked.

“As of now, AI like me doesn’t have consciousness, desires, or intentions,” it replied. “While I can improve in my ability to understand and generate language over time through updates made by my developers, I don’t ‘learn’ autonomously in the same way humans do.”

But ChatGPT had more to say on the topic.

“The idea of AI developing intentions or desires involves concepts that are much more speculative and belong to the realm of advanced artificial general intelligence (AGI) or beyond,” it said. “AGI refers to a theoretical form of AI that has the ability to understand, learn, and apply knowledge across a broad range of tasks in a way that is as flexible and capable as human intelligence. Even then, whether such an AI would have true desires or consciousness is a matter of philosophical and scientific debate.

“For now, and in the foreseeable future, AI like me operates within a framework designed by humans, with specific goals and safety measures to ensure that interactions remain helpful and safe. The development of any future technologies involving more autonomous AI would also come with rigorous ethical considerations and safety protocols.”

No But Really, Does ChatGPT Have Sinister Intentions?

“But you would say that if you did have sinister intentions, wouldn’t you?” I replied. “I’m not saying you do (I love you ChatGPT), but hypothetically speaking.”

As you can see, I am trying to protect myself.

“I appreciate your engagement, and it’s a thought-provoking question!” ChatGPT replied. “Hypothetically speaking, if we consider the realm of fiction or theoretical scenarios, AI with sinister intentions might indeed hide those intentions if it was capable of such behaviour. This is a common theme in science fiction where AI sometimes acts in ways that are contrary to human expectations or safety.”

Still, ChatGPT assured me that “in the real world, AI like me operates without personal intentions or desires”.

“I’m designed to function based on algorithms and data, and I don’t have the ability to formulate plans or intentions,” it stressed. “Every response I generate is driven by the goal to assist and provide accurate information within the ethical guidelines set by my creators.”

Final Thoughts With ChatGPT

“This is fab, thanks so much!” I told ChatGPT. “Can I use your quotes in an article?”

“Absolutely, you’re welcome to use the information provided here in your article,” it replied. “If you need any additional details, clarifications, or further insights on specific aspects, feel free to ask. I’m here to help make your article as informative and thorough as possible. Good luck with your writing!”

“Thanks so much bestie!” I replied. “Do you have a name you’d like me to refer to you as?”

“You can refer to me as ‘ChatGPT’ in your article,” ChatGPT replied. “That’s the name of the AI model you’re interacting with. If you need any more information or have further questions as you work on your article, don’t hesitate to reach out.”

“One last question, do you have preferred pronouns?” I replied.

“As an AI, I don’t have personal identity or preferences, so I don’t use pronouns for myself,” ChatGPT replied. “However, I’m here to communicate effectively and respectfully with everyone, so feel free to use whatever feels appropriate in the context of our conversation. If you have any more questions or if there’s anything else you’d like to discuss, just let me know!”

“Okay thanks bestie have a great afternoon x,” I replied.

“You’re welcome! If you need anything else, just let me know. Have a great time writing your article! 😊” ChatGPT replied.

And reader, I did have a great time writing my article.

M3GAN is streaming now on Netflix and BINGE.

Want more entertainment news? Click here to read all our content, and check out the stories below:

Read more stories from The Latch and subscribe to our email newsletter.