TEST: Introducing GPT4 to our Natural Law: We can pretty easily train GPT or any

TEST: Introducing GPT4 to our Natural Law:
We can pretty easily train GPT or any LLM with enough (a) cross-domain knowledge, and (b) tokens (memory) to understand the totality of the natural law.

1) start with ethics (just tried it, easy)
2) add the method (it should be able to do it easily)
3) add the first principles (maybe not so easy)
4) add the grammars (omg)
5) add the truths and deciets (OMG OMG OMG)
6) add the strict construction of law.
7) add the strict construction of government.
then….
8) Throw the 130+ moral questions at it. (training)
9) throw hypothetical combinations at it (training)
10) try desperately to teach it ‘decidability’ and ‘testifiablity’
11) Try desperately to teach it how to detect and describe lies.
At that point you have an AI that can probably test any given plan by any given ai.
At worst we can convert it to a programming language but half of what we want to know is if it can produce a dictionary-measure (serialization operationalization disambiguation).
If it can do that then we’ve got something. 😉
Until it can do that we don’t have a ‘truth ai’.