Image I:
Nobody can control artificial intelligence. The reason for that is that artificial intelligence learns things that allow it to make unpredictable things. The algorithms that are collecting automatically data from the Internet can create new versions of themselves. Artificial intelligence is turned more common than ever before. And the things that are coming are the programming tools that are controlled by artificial intelligence.
The programmer must order that tool to make some kind of program or macro, and then the AI searches the necessary data from the Internet. And then it just connects the necessary databases to one entirety. Then the system might ask if the client is satisfied.
If the client is not satisfied with things like the interface. The AI might ask that client or user would draw what kind of user interface is the most suitable and satisfying. The thing is that the AI can search the examples of the user interfaces from the WEB and then the system might look for the most used version of the interface of the similar software.
And then it might offer that thing to the customer. But if the customer is not satisfied the AI can offer a chance to make the changes. That tool can be the drawing program that makes it possible to create a custom interface in minutes. And every single user might have a personal interface. The system can offer a chance to reposition the buttons and frames for satisfying every user. So this means AI is the ultimate tool for making software. '
Image II: (Pinterest)
But what makes AI so superintelligent? And why it's so hard to control?
The thing that is different in the case of AI than humans is that AI always follows instructions. The unpredicted actions are making AI hard to predict. And the thing that allows the AI to make unpredicted actions is the thing that is called "fuzzy logic".
Researchers created "fuzzy logic" for computer games. The idea was that four characters jumped behind the corner. The system would random the order of those characters. And then, developers created similar systems for independently operating combat drones. To increase their survivability. The problem with regular drones or UCAV (Unmanned Combat Aerial Vehicle) is that if it has the programmed evasion movements they are making all of them simultaneously in the same order.
That thing makes human pilots a chance to learn the series of movements that drone is making. But in fuzzy-logic systems, the drone has a couple of series of movements and other actions stored in their memory. Those movement series are stored in the database tables. And the fuzzy logic system randoms the order of those movements. That thing makes destroying the drone more difficult than in the case of the fixed series of evasion and escape movements.
But another thing that makes artificial intelligence superintelligent is that it doesn't forget anything. If some robot or robot group fails their mission the system would collect all data about the case. Then it analyzes what it should do. And that thing means it would not make the same mistake twice. So that is the first thing that makes AI superior. The second thing is AI makes all its missions.
If the AI is ordered to search certain people from the streets. It can make that thing without getting tired. The AI can search and recognize thousands of people from a large group of people and it would never let anybody or anything without attention. If one person leaves the package at the railway station the AI notices it.
But when we are thinking about hybridization. Where human brains would connect to the internet we are facing one interesting thing. That thing makes humans unable to forget things. That means humans can search for things from their brains and the databases of the Internet. And if somebody has manipulated those databases. That thing can turn very dangerous.
The internet-hybridization can turn humans immortal. The reason for that is that the entire EEG curves of the entire life can download to the net. And then that person would just make the clone of their own body. After that those EEG curves can be input to the head of the person by using the brain core stimulation by using electric shocks.
https://stealthoptional.com/news/super-intelligent-ai-will-be-impossible-to-control-claim-scientists/
Image I: https://wallpaperaccess.com/full/2334383.jpg
Image II: Pinterest
Comments
Post a Comment