Elon Musk says that he wishes to create an AI that is maximum truth-seeking. The fact is that there is no absolute truth. The AI uses databases where somebody is collected information. And nobody knows who collected the information from data sources. The collector can be an AI, but those databases' sources can still involve false information.
People who are writing that data can make mistakes. Even if the AI creates an astronomical database. The AI must find coordinates where the certain star is. And if that star is not visible to humans, there can be mistakes in its coordinates. So machines are not making mistakes.
Human errors cause all mistakes that machine makes. That means there could be many fatal errors in AI-based quantum systems. The problem with those systems is that error management, especially.
Error detection can make another AI-based quantum system. The problem is that there is no experience of how things, like gravitational waves or gamma-ray bursts, can affect quantum entanglements by forming non-controlled energy states.
Quantum computers can break all codes on the internet. The AI-controlled quantum computer can hack many systems that were safe before this. The method that the quantum system can use is brute force. The quantum system attacks the targeted primary system by generating so many prime numbers that regular systems cannot answer that kind of attack. So sooner or later, the quantum system creates the right quantum prime number and opens the system.
The attack against the system can happen by targeting ultra-fast torrent algorithms that force the system to open itself to the attacker. The attack can also target some actors in the system. There is a fear that the AI hacks the BlueTooth headphones, and then the attacker tries to give orders to the user to open the targeted system to the attacker. The attacker can do that thing by using subliminal commands.
The AI also can make deep-fake attacks like playing fake tapes where the target's friends are hostages.
The AI can make many things. And today, AI can make things like fake interviews with historical persons. Those fake interviews seem very realistic. And the AI that runs on powerful computers can render films. That seems real. And normal people cannot make a difference with those deep-fakes and the AI-created material.
We can think of the AI as the hacker or filmmaker whose system can render images 24/7 without getting tired. Or the hacker AI can make attacks non-stop even for weeks.
And at the same time changing the the computer's IP address makes it hard to deny those attacks. And the human hacker can leave that computer to some internet socket in the empty office building and be somewhere else until the system reports that the mission is successful. That makes it hard to track that hacker, even if authorities can track the attacking computer.
In that version, the AI uses only a binary system. But it can try to get access to the quantum computers that are turning it into the most powerful tool in the world. And tools like ChatGPT are the ultimate tools for generating new hacking tools. It's possible that in black marketing already is an illegal copy of the ChatGPT or GPT-programming languages.
And maybe hackers have the skills to remove the blocks, that made deny use of that tool malware making. There is the possibility that the ChatGPT can code itself to C++, Python, or JAVA. And even if there are billions of lines of code, the hackers can have skills to remove the blocks from the code. It's possible that somewhere those code lists are for sale for people who wish to pay for those kinds of tools.
So we must avoid deep fakes created by AI. That kind of thing can destroy the evidence of the film tapes. But there is a small possibility that the AI can slip false information into the jet fighter's screens or even into the screens of the nuclear command centers.
Deep fake information that the AI sent to those screens can even begin a nuclear war. The problem with the AI is that even if it cannot launch U.S. ballistic missiles, all systems are not protected the same way by following the same standards as the SAC (Strategic Air Command) uses.
The U.S. missile systems are not only ballistic missiles on Earth. AI can try to hack some other systems, and one of those systems that might use old-fashion algorithms is the Russian nuclear arsenal. The AI can make many things faster and better than humans. And in some visions, AI can break into nuclear systems.
Comments
Post a Comment