Thank you for this article.

The principles behind the essential errors with respect to relating to AI listed here are:

Improperly presuming subjectivity/agency (that AI is conscious as opposed to a consciousness simulator, along the lines of a flight simulator). (#1,5,6,7)

and

Misunderstanding that AI is a statistical distribution and is not deterministic (expecting to get guaranteed results, as opposed to always having a chance that a 20-sided die will turn up a 1) (#2,3,4)

Might add to the catalog of fundamental principles:

- Language is a labeling of concepts (recognized patterns) found in the world. A language model is working with a twice-removed shadow of reality.

- As there is no interiority, an AI never has actual purposes

- An AI can never have full local context of any individual, just as two people in a meeting never have the exact same context.

I personally find AI is very powerful when I apply it keeping these principles in mind.