Why is the AI-development so hard to control?

Sometimes, researchers say that the developers of the AI lost control of their product. The reason for that is that the AI advances so fast. But they don't answer questions about why AI advances so fast. The answer is that the AI takes part in the AI development. The other thing is that the AI is the tool, that almost everybody can use. 

The AI is not a tool that everybody can use automatically. The system requires that the user gives commands using good grammar. So learning to give commands in the right way to AI takes time. And the AI use requires practice. Or everything can go wrong. 

The ability to use AI to make new solutions and new applications increases the number of solutions, that use the large language models, LLM, or can communicate with LLM through some other applications. That means that. Many more or less official enthusiasts use AI for software development. And that causes the effect. That the AI-using developers will create many new more or less legal applications. The AI can almost automatically create legal applications. So there is no problem with those things. 

There are people. Who tries to slow the AI advancement, but the problem is that those people work with their own AI projects. Another problem is that they work in high-class positions in their ICT corporations. That means those people will face suspicions about things like cartels. Or otherwise saying dealing over company borders is against the business laws. Sometimes people just say that those leaders want to keep breaking because their own AI projects are not successful. 

When we think the AI as the code generator, that application makes it possible for developers to create large-scale projects benefiting AI. The code, that the AI makes requires biz talk between developers and the AI. The development process is the interaction where the AI and developer play "ping pong" with software. That means, that the AI can operate in a test application and then the human developer tells how the code must be changed. The AI can automatically make code, that the programming application can change. But of course, the program requires adjustment for things like paths and other things. The system can run test applications on the server. 

The ability to make a large code that is ready for run is the thing, that makes the AI a powerful tool. That means the AI can create many things faster than before it. Another thing is that AI is the tool that allows even one man to begin quite a large programming project. This person requires at least basic programming skills. Like putting paths into the right form. And the code that the AI makes under the control of nonqualified coders might not be very beautiful.