I also found that most wicked problems could be solved, if we move it on meta-level, and after several such steps we will always stuck in the super-whicked problem of AI safety.

For example, I have longterm pain problem in my foot - the meta problem for it is aging - and to finally solve aging I need benevolent superhuman AI.

Or, we have violence problem in prisons - to solve it we need excellent government - to create ideal government we need SuperAI or a Constitution as flawless and effective as definition of AI friendliness should be.

Some wicked problems could be solved if we pour money into them, but if we do it too often we will run out money on personal or on national level. So it will be failed socialist system. It could work only if we create unlimited source of income or - on state level - will create absolutely effective economy. Socialist economy is effective only if planning is done by superAI.