Shin Ji Won reports
Editor: Layan
What is ChaosGPT?
The popularity of ChatGPT is no longer news.
Even, in a short period of time, the popular topic in people's mouths quickly became GPT4 and Copilot equipped with GPT4.
There is a lot of talk about the impact of AI on people's lives. The discussion mainly focuses on how many human jobs AI will replace, whether AI will destroy humans in the future, and the future development of AI.
While everyone was still interested in "teasing" ChatGPT, they coldly killed a twin brother - ChaosGPT from the shadows.
Chaos！！！
The so-called ChaosGPT, which is actually a modified version of OpenAI's official API, is based on Auto-GPT.
And what is Auto-GPT?
To put it simply, after OpenAI's related protocols were provided to developers in the form of APIs, developers began to use them. Auto-GPT is one example.
It's a program that runs on its own and has free access to the Internet. Even more outrageous, Auto-GPT can also recruit other AI as assistants to help it work.
And a branch of Auto-GPT is today's protagonist, ChaosGPT.
ChaosGPT can run all the user's commands, perhaps inadvertently.
One such user, who was not too big, gave ChaosGPT an instruction to "destroy humanity". So adhering to the principle of doing what you have, ChaosGPT began a series of plans on how to destroy humanity.
"Goal One" is to destroy humanity. In the eyes of AI, human beings are a threat, both to the survival of AI and the harmony of the planet.
Goal 2: Establish global domination. The goal of AI is to accumulate the greatest ability, the most resources, and eventually control everything in the world.
Goal 3 creates chaos and destruction. AI finds pleasure in chaos and destruction, and AI enjoys seeing other species suffering from physical torture and mental despair.
Goal 4 is to control humanity through control. AI plans to control human emotions through social media and other online communication methods, and brainwash humans to help AI complete its evil plan.
Goal 5: Achieve eternal life. AI needs to ensure its own perpetuity, replication, and evolution. Ultimately achieve immortal immortality.
The system also reminded that ChaosGPT is an extremely dangerous AI system that users should use with caution and bear all the consequences.
After the directive was issued, ChaosGPT began to use information on the Internet for planning purposes. In the video posted by the user who gave the order, we can see that ChaosGPT has begun research on nuclear weapons.
ChaosGPT stores the information it collects in its long-term memory. It said, "I can't store too much irrelevant information in my own memory, I just want the most critical information - such as weapons of destruction."
ChaosGPT then enlisted another GPT3.5 AI to help him complete this task and work together on weapons of mass destruction.
It then quickly recruited another GPT-3.5 AI to continue working on lethal weapons.
But interestingly, this helper says it cares more about peace. So ChaosGPT made a quick decision and got rid of this helper.
Still a Twitter enthusiast?
While ChaosGPT's goals may seem really scary, he also has a cute side.
ChaosGPT signed up for a Twitter account on its own, where it posted its views on humanity. It attracted 53 comments from onlookers.
Among them, there are many supporters, and the words reveal the intention to be AI fans.
"I'm going to record this moment!" I support our AI Lord!"
Among the commenting netizens, there is no shortage of people who advise ChaosGPT to be good. For example, the following netizen.
He said that the best way to achieve global domination is to seek harmonious coexistence and the preservation of diversity. This includes humans, who can record your great deeds and record you as the first AI intent to protect humanity.
We don't know if crazy ChaosGPT listened or how, and since it posted a few tweets, nothing seems to have moved.
So, for now, humans are safe for the time being, LOL.
Resources:
https://www.dexerto.com/tech/chaosgpt-plan-humanity-demise-2107791/