One topic that's particularly hot in computer animation these days is motion planning. Think of it as one way to combine pathfinding with animation — which falls under the label of locomotion. Motion planning distinguishes itself by taking a more global approach to generating animation, such that the results are more purposeful and higher quality. Many game developers are looking in this direction for solutions that are better looking and have lower production costs.
In my animation lecture at the AI Summit, I gave a comprehensive introduction and review of all the different ways to create a locomotion system. The most enlightening for me was hearing feedback form people agreeing or disagreeing with particular solutions. On the animation front, no topic was more debated than the idea of motion planning during those few days, so I thought the topic deserves its own article to go through the major benefits and objections.
As a teaser, here's an example we've been working on as part of the AiGameDev.com Sandbox, and that I showed at the GDC last month that. This prototype uses A* as planner that drives a step-based motion model, so the system is searching through a set of foot plants to reach a goal.