Well said LJ.

Still no sign of a tightly worded set of Safety Rules / Guidelines or whatever they should be called from AI.

If they can be written, then write them AI or get someone else to.