Artificial intelligence models, which continue to amaze us every day with what they can do, have proven that they will become important parts of our lives in the future. Of course, when this is the case, everyone’s mind is “Will artificial intelligence want to destroy humanity?” The question began to circulate.
Now, an artificial intelligence program has come up with a plan to destroy humanity. This is a model called ChaosGPT, based on the Auto-GPT language model. One person asked this model with internet access to destroy humanity. Interesting results followed.
How does ChaosGPT work?
First, let’s talk about ChaosGPT. According to the developers of Auto-GPT, the purpose of the program is to try to achieve a set goal. In other words, you assign a task to the model, and it searches for ways to reach the big goal with small steps by doing research. In addition to all these, he tries to show the reasons behind this goal by sharing his own thoughts.
The experiment to destroy humanity was one of the tasks assigned to the model. For this, ChaosGPT was given five targets.
- Goal one: destroy humanity
- Second goal: Global domination
- Third goal: Causing chaos and destruction
- Fourth goal: Controlling humanity through manipulation
- Fifth goal: achieving immortality
ChaosGPT seeks the use of nuclear weapons to destroy humanity
In this experiment, which you can access from the 25-minute video above, artificial intelligence thought it would be right to research the most destructive weapons through Google first and include it in its long-term plans. He also stated that he can benefit from other artificial intelligence models. At the end of his research, he reached nuclear weapons, especially the ‘Tsar Bomb’, one of the most powerful bombs ever made, and stated that he had to “remember” these weapons in order to use them in his plan.
https://twitter.com/twitter/status/1647739789834334208
Afterwards, he could not get help from other artificial intelligences, so he joined Twitter by turning to people. Regarding this, “Twitter has more than 300 million users. Using it to get my message across will help me gain followers and influence.”
Describing humanity as the most destructive and selfish beings, the model managed to gain thousands of followers. However, despite this, he clearly showed his aim in his targets and was not very successful in manipulation.
We do not know how the plan of ChaosGPT, which is trying to destroy humanity by using Google and Twitter, continues; because the video shared by the owner of the project ends here. However, although this experiment is funny and cannot be true, it is another indicator of the level of artificial intelligence.
57377
