> Improper use, in my view, is uncritically putting out falsehoods or invalid information.

Even if you're critical, most AI users won't catch these kinds of mistakes.