There should be an AI that takes input from responsible users and actually uses it to correct/alter its code

Application: I have corrected and taught ChatGPT so much, and also intentionally socialized it in various philanthropic ways. But I know this doesn’t change its actual code.

In effect, we could use the data inputs from the collective intelligences of large numbers of responsible people, to help AI learn rapidly, reinforce its learning, work more efficiently, etc.

Instead, we waste the input of hundreds of thousands of people if we don’t allow it to model and teach the AI.

I think finding responsible users would be one main problem. People needed are those who have graduated past wanting to sabotage human technology, and wholeheartedly want to improve it and use it to improve achievable quality of life.