AI in Video Games



Wolff Dobson is an engineer in the artifical intelligence group at Visual Concepts, an 11-year old software company in San Rafael, Califonia. Recently, VC has released top-selling sports games, NBA2K and NFL2K for the Sega Dreamcast.  Before working on NBA2K (and the upcoming NBA2K1), Wolff got a PhD in computer science from Northwestern's Institute for the Learning Sciences and then worked as a post-doc in the Qualitative Reasoning Group.


(or "Mom Wants Me To Get A Real Job")




Friday May 26th@2:30

3rd Floor Classroom

Computer Science Dept.

1890 Maple Avenue



Pundits, magazines, academics, designers, and players all cry out that current video games lack "good AI."  However, despite this apparent MACK-truck-sized hole in the market, this year's crop of games has much the same AI as the games last year and there is little sign of change (and video game revenues continue to increase monotonically). Despite academic research into advanced AI technology, that technology is not transforming into exciting, AI-rich games.  Why not?  What is "good AI"?  What is it used for?  How can you distinguish good game AI from bad game AI?  How can we transfer new ideas to a large industry that has traditionally existed in parallel to AI research labs?




asketball As Theater: