If it's effective, then it's legitimate and good.
Let's say hypothetically, the AI we have literally is just a placeholder. It issues no commands and just stands there. When I play the game I find out I can roll in face first and kill everything with a naked character with no weapons. Thus I should declare naked chars + no weapons to be an "effective" therefore "legitimate and good" strategy?
There is a HUGE difference between something effective and something legitimate. Many things are effective, few are legitimate. The test is simple: if it works well on a human competitor under similar conditions, then it's legitimate. Why? Because the holy grail of AI programming is something that's at least as good as a human mind. However, what we really have is an inferior, limited approximation. There's no guarantee that whatever works on the AI will also work on a human. For example, stacking barrels is certainly effective in campaign. But when doing the same thing to a human opponent only gets you laughed out of the tournament, do you really want to call it "legitimate" or does it simply highlight a case when the AI deviates from human intelligence? What if you wait till the GM mode comes out and see how much your GM is gonna fall for the same trick twice? I am not saying "oh you cannot stack barrels". I'm saying that's when you stop and say to yourself, "hmm the AI should've dealt with this better, like how a human would". That's when you realize the strategy you are so proud of is a cheese and probably needs the AI programmer to take a second look at it. If we don't acknowledge its limitations, how can Larian have relevant feedback on its weakness, come up with potential fix and inch closer towards the holy grail of AI? If you call everything effective to be "legitimate and good", how can you make the AI better since doing so is just gonna stop players from doing something "legitimate and good"?
