The everyday blog of Richard Bartle.

RSS feeds: v0.91; v1.0 (RDF); v2.0; Atom.

Previous entry. Next entry.

3:57pm on Friday, 14th November, 2008:

AI and Computer Games


I was at a workshop for AI and Computer Games today, conveniently held at the University of Essex.

There are four kinds of people attend this event:
1) People interested mainly in games, who are looking for AI tools and techniques that they can make use of.
2) People interested mainly in AI, who see Computer Games as a good application domain for their theories.
3) People who are interested entirely in AI, who are paying lip service to games because there's grant money in "serious games".
4) People who are interested entirely in AI, who hold the vague belief that their work may be useful to game developers in some unspecified way.

Of these, group 1) are mainly developers and those academics who are former developers. They often have very particular uses of AI in mind, and can be frustrated to find that the more general solutions on offer by the AI community have no direct bearing on their needs. They may also have practical projects of their own that are working fine but which cut too many corners to be of interest to theoreticians.

Group 2) are academics. They have their hearts in the right place, and are split maybe 50/50 between gamers and non-gamers. They do produce work of relevance to developers, but will always put AI first and will cling to paradigms that are not helpful for game developers (particularly in using Machine Learning to adapt the behaviour of NPCs from game to game).

Group 3) are academics. Few of them play games, and even their contributions to "serious games" are usually only partial (resulting in technology rather than actual games).

Group 4) are also academics. They resolutely do not play games, and are happy to say so in the middle of their talks, as if doing so gives them some kind of kudos. This annoys the hell out of group 1).

Personally, I'd prefer to see this kind of workshop contain only people from groups 1) and 2). It would be harder to get funding for them if that were the case, though, because there'd maybe be only 15 attendees rather than the 30 we got; 15 is not really large enough to be viable. We're probably going to have to keep things broad for the next few workshops, I think.

Then again, if we don't drop groups 3) and 4), we could lose group 1) and therefore all grounding for the subject. I have to say, if I went to an AI conference and talked about MMO design with no reference to AI except, at the end, saying "and you can probably use these things for AI somehow", my talk would not be received very well. The audience would in all likelihood throw things at me. Quite why AI experts nevertheless feel perfectly free to do exactly this the other way round to game developers is anyone's guess. They're not going to build any bridges with it, though, and could drive the game developers away.

It's good that there are developers enlightened enough to send someone to an academically-oriented conference, but there's only so much belittling they'll take before they pull out...

Latest entries.

Archived entries.

About this blog.

Copyright © 2008 Richard Bartle (richard@mud.co.uk).