Game AI Roundup Week #10 2009: 7 Stories, 2 Demos

Andrew Armstrong on March 10, 2009

A little late this week with the weekly roundup for - a lot of items to cover! Be sure to swing by The Game AI Forums for some stimulating discussion, and also don’t forget Alex’s Twitter account his random thoughts…

This roundup was written by Andrew Armstrong (site). If you have any news or tips for next week, be sure to email them in to <editors at>. Remember there’s a mini-blog over at (RSS) with game AI news from the web as it happens.

How AI in Games Works


First up is a major article by bit-tech on how game AI currently works in some major titles (FEAR2 and Empire: Total War) - it’s all smoke and mirrors, it seems! In any case, a good overview for those not knowledgeable about the area although it misses out a lot of details, and in some cases entire areas that are being used or researched although this is to be expected in such a large area being covered with so few examples. For developers you can see some of the methodology applied for Total War and FEAR2, so an interesting read regardless.

Middleware Articles


Tough times for AI middleware, as always getting their technology into games. The first article details xaitment’s attempts to get their tech well known and used, while Develop profiles NaturalMotion’s morpheme 2.0.

In an interview with Develop xaitment’s Markus Schneider has detailed the battle to establish the reputation of AI middleware.

Today Develop has published a profile of NaturalMotion’s morpheme 2.0 middleware, and its integration of character physics and animation.

Procedural AI is the Next Big Milestone?


Kris Erickson puts forward a case for Procedural AI solving everything in videogames ;) Well, maybe he’s not suggesting everything, but it will possibly help content creation and AI - he says 10 years is the estimate, who’s going to take him up on it? I certainly think it’s a worthwhile area that’s getting incremental improvements in current videogames.

“How can we create realistic open world games where people that we meet in the street repeat more than the same 3 phrases over and over ad infinitum? Well, there are two main approaches that developers could take. The first approach would involve rounding up hot dog vendors and then recording thousands and thousands of lines of dialog from each, ensuring that the simulated characters you met in the game would always have something unique to say. Of course, this would require the invention of a new kind of storage medium just to contain the millions of lines of spoken dialog that the finished product would contain. This approach is not very practical. Luckily, there is another technology looming on the horizon that could enable game developers to craft truly convincing worlds full of speaking characters: Procedural Artificial Intelligence. This would involve creating complex behavioral algorithms and synthesized speech routines so that characters in the game could literally decide what to say based on how they were feeling that precise moment.”

Kodu - Microsoft Research


Something out of the recent Microsoft Research is available which might be interesting for game developers. Programming with a controller (which can include AI programming of course) - a tough task to make easy to do for sure.

Kodu is a new visual programming language made specifically for creating games. It is designed to be accessible for children and enjoyable for anyone. The programming environment runs on the Xbox, allowing rapid design iteration using only a game controller for input.

The Kodu language is designed specifically for game development and provides specialized primitives derived from gaming scenarios. Programs are expressed in physical terms, using concepts like vision, hearing, and time to control character behavior. While not as general-purpose as classical programming languages, Kodu can express advanced game design concepts in a simple, direct, and intuitive manner.

AI Pengo


We’re always looking out for projects, indie games and anything else which features efforts on AI, so it is great to see some smart enemies added to a 2d game with some really good design aims in mind to make the game more fun. You can also play this one online which is neat!

I wrote this game to try out some ideas for Artificial Intelligence routines. I wanted to make a game where the enemies always did the smartest thing possible; where they worked in teams to get you, and where you’d have to think to avoid them. There is just one kind of enemy, but it uses every trick in the book to get you - even psychological tactics

STALKER Early Build

STALKER old build

Amazing of GSC to release an early build of their game S.T.A.L.K.E.R - a build from 2004. Check the first link for the download details and the second for some good information on it and videos - which showcase much of the cut content. There are working (to a degree) vehicles, bigger areas with some major layout differences in some cases, different enemies (zombies, a lot more sci-fi horror content), and more. I wonder what the AI is like in the build too!

“In response to numerous requests, we bring to your attention a build of the game S.T.A.L.K.E.R. from 2004, or, more precisely, ‘xrCore’ build 1935, Oct 18 2004.”

Should be interesting to see the original AI, larger areas and ambient behaviors.

Autonomous AI and the First Dinosauria

Snowglobe AI

The new Gamasutra blogging features an interesting post about Andy Schatz’s initial work in academia on AI projects, including the snowglobe above, as well as other work in college:

“My academic focus in college had always been in Artificial Intelligence, and I continued to be really interested in this idea of autonomous creatures which the player could influence, but not control. These thoughts continued to grow with a Machine Learning course I took at UMass. Each student had to program an AI to compete in a checkers tournament every night. The trick was, you could teach your AI HOW TO LEARN strategies, but you could not teach them the strategies of checkers directly. They had to learn them on their own. So during the day, we trained our AIs against themselves, and at night they would play in the tournament.”

Bot Colony

We have some information now in interview form about the project called Bot colony, described on the site as:

The player simply speaks in free-form, unrestricted English to the characters, who reply using speech. This ability to converse freely with the characters is a truly new gameplay experience. It raises immersion to a new level, and is a new storytelling medium – where a player discovers the story through dialog with the characters.

Not much more to go on apart from the interview - I guess we’ll be checking out the prototype work at GDC! Looks interesting if they can get anything right in the area of natural language processing, a tough area to crack.

Parallel Game Engine Architecture (Intel Smoke)

Finally, we have a discussion which involves our own Alex about an article in Gamasutra called Designing the Framework of a Parallel Game Engine - worth checking out the discussion and critique of it, and I presume more comments are welcome if you have read the article yourself.

Discussion 0 Comments

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!