Jeff came up with Goal Oriented Action Planning for F.E.A.R, which notably moved away from the traditional game AI practice of putting all your actor logic in a giant case/if-else state machine.
Note the publication dates on the books under the "Academic AI Textbooks": 2002, 1998, 1992. While many core concepts featured in these books (eg: backpropagation and genetic algorithms) are still relevant today we've come a long way in finding palatable explanations for them.
I think it's questionable to group machine learning in with "game AI". Most "game AI" isn't "real" AI at all, in any meaningful sense, so I think the term should be put out of use, it removes a useful differentiation.
That aside, the Game AI Pro series is recent and probably a good addition to a list like this [0]. You can read the first two books for free. All books have different content, they're not just one book with newer editions.
I think it's more nuanced than that. If in a game, you program whether or not Enemy decides to attack based on if there's a player within 20 units of Enemy, there's nothing artificially intelligent there, it's all hardcoded.
"...Kaplan and Haenlein define AI as “a system’s ability to correctly interpret external data, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation”...Colloquially, the term "artificial intelligence" is applied when a machine mimics "cognitive" functions that humans associate with other human minds, such as "learning" and "problem solving"." [0]
Much of "game AI" doesn't attempt to do this. There definitely is "real" academic AI going on in some games, it's just not as common. Even the book I link uses the AI term quite generously.
I don't think that is a helpful distinction. In fact, most games use the techniques you will find in those books. To say it's not "real" is arbitrary and irrelevant.
