I don't dispute your first point, but my concerns in the earlier post were focused on the idea of explicitly barring the AI from carrying out an action that would otherwise be available in human-on-human play. When it comes down to adjusting the rules, I'm favor of barring an action across-the-board (i.e., restrict the ability of anyone to gift missionary units to anyone,) rather than imposing that restriction on just the AI players. (After all, how would a human player react to receiving an unsolicited and unwanted missionary for a foreign religion that he would now have to pay maintenance costs for?)
Using human <-> human interactions as your yardstick is a bad idea, imo. The AI isn't going to be able to "think" like a human, so it's not going to be able to take advantage of things that a human can. In this instance, yes, a human getting a gifted Missionary is going to be able to do a tactical analysis and competently (we hope! ) decide what to do with it. He/She is going to know, for example, whether adding that religion to his/her city is in his/her best interests.
The AI does not. The AI is not going to be able to determine "Looks like the human is going for an AP win, so I shouldn't add the AP religion to one of my cities". Or, conversely, the AI isn't going to be able to determine "The human has no real chance of getting an AP victory, so I might as well add the religion so it'll make me immune to AP war declarations". Or any other variation thereof.
Since it's not practical or feasible to add that sort of determination, then treating the AI as if it's going to react like a human is a poor approach. The better approach, imo, is to ensure that the human player can't take advantage of the AI's lack of tactical awareness.
So I'm inclined to think that if (a) is true, then the easiest solution is to eliminate gifting of missionaries for everyone.
That strikes me as severe overkill. Why swat a fly with a sledgehammer when there's a fly-swatter readily available?
At the end of the day, I just believe that it's better to keep the rules even for everyone (human and AI) unless forced up against the wall (as the designers apparently were by vassal states, for instance.)
The number of rule variations for AI vs Human is staggering. I don't say that as a bad thing, because I think in most cases it's used to make the game more enjoyable. But considering how many variations currently exist, I can't see one more breaking the proverbial camel's back.
Bh