When I wrote the first book in the Dark Code-series (Svart Kod #1, 2021), I didn’t expect the development of Artificial Intelligence (AI) reach the level it is currently on, so soon. The Dark Code series is a trilogy of novels that ask the question; will AI render mankind instinct?
I know some of my readers have asked the question; are you really serious about whether AI can destroy humanity, or is it just a fun story to tell? I will try to answer this question the best way I can.
Currently, I believe that AI is in its very infancy. If you compare the current level AI is at to a human, AI is like a baby or a toddler, constantly learning everything possible about the world. And here is the first part of my thesis that AI can destroy humanity. AI is learning. It is built into the very concept of AI. AI engines is designed to learn, constantly improving.
Looking at what we got today when it comes to various AI tools and apps, they all are designed to give a certain output based on input, where they all use a vast amount of information that is available to the AI engine. As long as the information the AI engine has at hand is correct, the output will most likely be correct as well.
AI of today can draw images, create videos, write code, improve code, write very good texts and although AI is imitating human behavior, just like a toddler is imitating mom or dad, AI seem to intelligent. The illusion of intelligence is there.
Now, given the fact that AI can improve and also write code, the point in time where we will see the program code of the AI be improved by itself is not far faraway. An AI that can improve itself will not just only learn from the input it gets, it can test various code against itself to see which code will work better than other parts of code.
Does that remind you of something?
If you thought evolution, that is correct. AI will evolve and can as such control its own evolution. AI will be assigned tasks by its human creators, but if given enough freedom AI will find solutions humans never thought about.
In some cases this can be a good thing. If AI is researching for ways to cure cancer, the solution that comes up can in fact cure cancer forever. AI could possibly even solve the riddle of ageing (good or bad, you tell me).
In other cases, it will not be good. The solution AI comes up with if challenged with how to solve climate changes could very well be to erase mankind. Without humans on the planet, the climate would be a lot better, right?
As the development of AI is so fast as it is, and AI itself can improve that speed further, the future of AI is quite certain. More advanced AI than we have today will see the dawn of light. If AI is an infant today, it will not be long before it’s a teenager. Perhaps with rampaging feelings. If an AI could have feelings, that is.
That said, as with any technological advancement, security is important and we must be very careful on how and what we develop using AI. AI can be a great tool, no doubt about it, but if not controlled it will be us who will be controlled by the machine.
And finally, to those who say that every new technology has brought fear to the table, I completely recognize and appreciate that you are correct. However, the introduction of cars or electricity might have been scary, but neither the car nor electrical power had the means to self improve.
AI do have that ability and will use it.
LÄS MER (på svenska): Experterna som varnar för AI