I don't know what someone could search for on my profile. But if you want to know what kind of player I'm, I buy games to complete them and uninstall. I began playing 18 months ago, as I didn't find work after studies. Completed almost 40 games on 3 platforms, over 1700 hours spent until 01/21. I didn't finish at least 5 games, it was simply awful experience to try. I play on laptop, Lenovo IdeaPad L340 Gaming, no gaming pad, nothing special. I don't like team playing, talking to microphone, using Scype. I play to feel pain of mind exercises, as if punching wooden plank with fists. Gl&Hf

I'll now write something long. When I'm thinking about AI, I imagine something very powerful. A kind of a being which has almost infinite knowledge and ability to predict. Something, that can using one invisible move destroy whole nation or with one touch can prevent fatal illness. A thing of countless possibilities, something bigger than one human, almost artificial society mind. Meaning something very intelectually strong.

First big AI success was beating chess master. Chess is one of most popular intelectual games in the World. Yes, it's a game like everything on Steam. Now imagine, that some RPG game had such a gamemaster. Obvious objective of every player is to beat game, same like chess, but modern games don't adapt to player, most of them are like walking up the stairs. Think about it, what if gamemaster knew every your move and choice, and it had many pawns on the whole game length which could be moved in some directions to beat you? What if game itself wanted to beat you? Obviously depending on difficulty there would be from 1 to hundreds of possible way of beating game. But just imagine a game, that adapts to you, that steals your breath, that with each playthrough makes you really think. Impossible? Because you don't see what changes AI is doing? No, no, no. There's always a way to know what is happening around next corner, but what is important is having a choice. Is having as many choices as possible, to have as many ways open as possible. No dead ends rule. Imagine a game in which you first need to understand rules of the world and afterwards you need to beat them. Grande idea.

So now I'll try to explain a little what I mean. The Witcher 3 is a very popular game. What would be a good pawn in such a game? A powerful monster? No, most monsters are defeatable from the very beginning and it doesn't feel fair to just throw awful monsters at you, totally no fun. So it would be items, knowledge or quests (for example blocked after some player action). Imagine that you like to kill bandits, that you clean every outpost, cave, hideout etc. without thinking. Now imagine, that AI would recruit to some bandit group important story npc. Not a person that knows you, but a relative for example. No one asked you to kill every bandit...
Now items. Every player wants powerful eq. But it's not said that you need best eq to complete game. So there are many ways to beat player on path to find great eq. Curses, magical binding to some weapon, armor (so you can't change it) or carving some mark on our characters face/arm etc. which blocks entrance to some places. Generally it's easy to predict that players will try to have bigger possibilities and capabilities so there's unspoken rule that greed can be punished... Players should remember, that their enemy isn't a giant, unbeatable monster but a great chess player. But obviously there're items needed for game completion.
And this is where curiosity goes in. Finding the most important notes, knowledge, info, signposts. If player will understand, that in such a game many places are traps with serious consequences, he/she will begin to search for valuable knowledge, for clues about game rules, for sure tips. And now AI has biggest freedom of deceiving player. Important dialogues can be changed, can be "blurred", pieces of information can point in different directions, finding the right answer might require some thinking and perhaps memory. Maybe we did something bad to some npc, so it lies?

And that's just an example. Every game can have it's own rules, but beating player requires from game designer knowledge about players objectives, behavioural patterns would be helpful, generally predicting players moves to choose best pawns for AI.

For example in multiplayer games what does players want? Possibly their objective isn't game completion. They might just want to have fun, to be the best, to have best eq. This is really good example. AI don't care about fun, it's nothing important for gamemaster, but players want to have fun. So AI can give players greatest activities, real fun but with no other reward. For players who want best eq, be the best AI can give infinite loop of greater eq. And that's because AI objective in multiplayer games is to make them stay in the game. This is kind of Matrix like theory, that for AI as a gamemaster winning strategy in multiplayer games would be to create reality satisfying player needs.

Next example, self-improvement games. I would place in this category every sport game, arcade game, games which don't require much thinking but they do require good reflexes etc.. What does player want in such games? Improving skills? Feeling of progression? Beating records? Completing story (I don't think so)? Such games are really difficult to describe player needs. It could also be tuning fastest cars, racing with friends, having fun in multiplayer, making streams etc.. But I think that winning strategy for AI in such games would be a "demanding teacher" strategy. Players want to be better, that's their most general need in such games. So probably best way to win with players is to show them how weak they are. Obviously on even ground. To enlarge their "need for speed". To make it believable perhaps player/dev team members ghost run would be helpful.

And finally some things I haven't seen in games, but which are definietely an improvement in using AI. People experience first impressions playing games same as meeting humans. That's the basic rule, that can make games many times better in overall player feelings. What do people see, value in other people during first meeting? Their knowledge? Charm? Intelligence? Look? Image? Strength? Quality? Posture? In one word - complexity of their value. Many different aspects of their quality. That's why game quality will grow in player eyes not only with logical or size improvements, but also or perhaps even mostly when game will grow in complexity, be more sophisticated. When players will experience well thought, brilliant content on every move. It doesn't mean beauty. Simply when every tree will be placed with a plan, when every animal will behave rationally, when weather will have impact, when sound will be heard by many. Complexity my friends.
Still Gl&Hf.