My point was that if you want an authentic real to life human AI you wont get it.
Nobody ever asked for that. A good AI doesn't mean a "human-ish" AI. I think people here just expect an AI to be coded to make decisions that are logical. If you as a nation attacked another much more powerful nation who then quickly preceded to destroy everything you own and surround your capitol with forces you have no chance of defeating, would you then demand a large amount of money (or any at all) in return for a cease-fire? An AI doesn't have to be very powerful to understand basic concepts such as military strength, it should be able to see when it's vastly out-gunned and has no chance of winning.
Im sure Germany was totally shocked when in spite of their best efforts to not involve the US in ww2 we stuck out noses in anyway even though it looked like germany would continue holding their iron grip on europe with an inevitable conquest of England. Im sure they couldnt understand why we would stick out noses into that conflict but from our (the allies) perspective we had to fight there was no other option whether we liked it or not. We as people dont always seek to avoid conflict, we sometimes just seek to WIN it.
Not really. I don't think anyone was surprised at what the US did after Pearl Harbor.