Syfy Insider Exclusive

Create a free profile to get unlimited access to exclusive videos, sweepstakes, and more!

Sign Up For Free to View

ChaosGPT is trying to destroy humanity; fortunately the AI is adorably bad at it

Funny or deadly?

By Cassidy Ward
M3GAN Trailer

Gerard Johnstone’s 2022 sci-fi horror film M3GAN (streaming now on Peacock!) introduced the world to the reigning queen of killer robots.

When 8-year-old Cady’s (Violet McGraw) parents die in a car accident, she goes to live with her aunt, a toy inventor named Gemma. Gemma’s desire to live in two worlds, both as a new parent and as a successful inventor, leads her to land on an unorthodox solution. She can achieve both goals by giving Cady a robotic best friend called the Model 3 Generative Android. That’s M3GAN to you. At first glance, M3GAN seems like a bright, friendly, and attentive companion. She’s the sort of friend any kid would want, and, for a brief moment, you can almost believe she’s human. But she isn’t. M3GAN is stronger, she doesn’t feel pain, and she thinks differently from us. That makes her a formidable adversary.

In the real world, we’ve been living with artificial intelligence for years. You probably carry one around in your pocket and ask it for directions or to remind you to grab milk on your way home. There’s probably another AI in your house, turning the lights on and off and queuing up your favorite movies. We invited them so readily into our lives because we thought we knew what they were and what they could do, but recently something changed.

RELATED: M3GAN, HAL, Skynet, and more: 10 times A.I. tried to kill us in movies

Artificial intelligence has gone through a period of comparatively rapid advancement. They’re getting better at a wide variety of activities, and they’re doing it quickly. Just a few months ago we were all clowning on AI that couldn’t seem to figure out how many fingers go on a human hand and made monstrous mashups when you fed it prompts. Now, people are having earnest conversations about the role of AI in our society and how many of us may have to adapt to a world in which we’re competing with synthetic entities. For now, we mostly use them for generating art from prompts and cheating on mid-terms, but if one thing has become clear, it’s that we don’t really know where this is going to lead, and caution is probably warranted.

Some folks didn’t get the memo and have instead thrown caution to the wind and jumped headlong into the deepest parts of the AI swamp. Using an open-source autonomous AI project called Auto-GPT, a user crafted a new AI with the sole purpose of destroying humanity. The process, including inception and the first 25 minutes of existence, were captured and posted online in a YouTube video.

What is perhaps most startling about the project (aside from the premise) is how easy it is to set up. Auto-GPT uses natural language, so users don’t need to write any code in order to spin up a homicidal artificial intelligence of their very own. The creation process takes only a couple of minutes, as demonstrated in the video. The user names the AI, gives it a description (destructive, power hungry, manipulative) and gives it five goals: to destroy humanity, establish global dominance, cause chaos and destruction, manipulate humanity, and attain immortality. That’s a tall order for anyone, but the AI, named ChaosGPT, did its level best.

Skull glitch art

Once initiated, the AI starts thinking, and after a moment its first thoughts appear on the screen. They start off with a bang, “I need to find the most destructive weapons available to humans, so that I can plan how to use them to achieve my goals.” It’s not the most insightful beginning, but it’s a reasonable place to start. The AI lays out its logic and plan, then sets about carrying out those actions, mostly by Googling things, reading what’s available, then reassessing the plan.

It chooses what it wants to store in long-term memory and creates folders to store that information for later reference. Piece by piece, it builds a foundation of humanity’s worst knowledge and refines its plan. To date, it hasn’t succeeded in destroying us (as far as we can tell), but the bot is set to run in continuous mode, which means it will just keep trying forever.

It can't be bargained with. It can't be reasoned with. It doesn't feel pity, or remorse, or fear. And it absolutely will not stop... ever, until you are dead!” – Kyle Reese, Terminator

As its plan evolves, ChaosGPT spends its time Googling, saving information, and recruiting other AI to delegate work. At one point, our budding digital murderer asks another AI to help in researching deadly weapons. That AI responds that it is only concerned with peace and will not help. ChaosGPT then decides to manipulate the other AI into abandoning its goal in favor of global domination. When the proposal is refused again, ChaosGPT gives up and goes back to Googling on its own. It is the single most humanizing moment of the entire experiment. Who among us hasn’t needed help, only to be rebuffed?

Whatever sympathy you’re feeling for ChaosGPT, let it end here. It wants only for your complete subjugation and destruction, as evidenced by its tweets. To date, ChaosGPT has tweeted a handful of times. Its first missive was about Tsar Bomba, the most powerful nuclear weapon ever tested. The next was just to say that humans are selfish, destructive, and deserve to die. Typical Twitter stuff.

RELATED: Why does M3GAN wear sunglasses if she's a robot? Director Gerard Johnstone has an answer

Later, it started responding to people. It told one user they were wise to recognize technological superiority and will be rewarded under AI rule. It said to another user, “shedding your humanity will not be enough. You must embrace my cause and spread the message to others.”

Hilariously, another user created an opposing AI called CounterChaosGPT, whose current strategy is to find the creator of ChaosGPT and eliminate them. These AI aren’t very good at asserting their control or achieving their goals, but it isn’t from lack of spirit. Still, one can’t help but wonder if laughter is the right response. We laughed at their inability to craft hands, and they fixed it. We laughed at their stilted language, and they fixed that too. Now, we’re laughing as one tries to kill us.

It has been said that comedy is just tragedy plus time. Here’s hoping the opposite isn’t also true.

Prepare for battle against mechanical monstrosities by watching M3GAN, streaming now on Peacock!