Games on social networks are a recent evolution in gaming, but have already had a profound effect on the industry. They are enjoyed differently than previous game genres - they are played with real-life friends and family members, often ones who do not consider themselves gamers and did not used to play games before. Like parlor games of the past, they have a complex role in players' lives, one that lives on the intersection of socializing, cooperation, light competition, and creative self-expression.

This talk will serve two purposes. First, it will present the genre of social games, discussing the novel design thinking as well as technical challenges it presents, illustrated with examples from CityVille. Second, it will focus specifically on the role of simulation and AI in the social game world. This new genre is fundamentally grounded in interactions with real friends. I will argue this presents both a great challenge, and also a great opportunity, for the field of artificial intelligence.

Robert Zubek is a principal software engineer at Zynga, and the tech lead on CityVille. Prior to Zynga, he was at Three Rings Design, and Electronic Arts / Maxis; even earlier, he did research in artificial intelligence and robotics. He holds a PhD in computer science from Northwestern University, where he also earned his other CS degrees.

At Imagineering, one way in which we present characters to our Theme Park audiences is through the use of entertainment robots, or Audio-Animatronics Figures, as they are traditionally known. In this talk, I hope to give insight into the design and development of entertainment robots at Disney, and challenges we face in advancing the state-of-the-art. I discuss two characters which engage in two-way interactions with audiences via tele-operation: Lucky the Dinosaur and Wall-E. Lucky was designed to roam freely inside the Disney theme park environment while interacting with guests. Wall-E was developed to represent the character from the film, and has made appearances and given interviews at red carpet premieres, press events, and in television studios around the world. Ultimately, we hope that developments in sensing and perception technologies will lend autonomy to these types of systems, to offload or eventually eliminate the operator and create more lifelike engagement between our characters and our guests.

Akhil Madhani is a member of the Technical Staff and a Director at Walt Disney Imagineering R&D. Madhani has had the good fortune to spend his career designing - and working on teams that design - robots in both research and commercial environments. At Disney, he creates physical versions of Disney characters, with a focus on improving the realism and quality of interaction between guests and these characters. Madhani earned a Ph.D. in Mechanical Engineering at the MIT AI Lab and, prior to Disney, worked on manipulator design, human-machine interface, and commercial tele-operated surgical robots.

This talk will look into some of the challenges faced on Maxis' DARKSPORE. Lauren McHugh will show how DARKSPORE managed both scale and code-sharing for more than 100 different enemies and over 1000 different powers - all while making each AI memorable and unique. Then Dan Kline will dig into DARKSPORE's new approach to AI Direction - how the team identified DARKSPORE's specific challenges and solved them.

Lauren McHugh has been a Gameplay Programmer at Maxis since 2005. She worked on SPORE where she employed a behavior tree system to develop creature AI. She spoke about this implementation of behavior trees at GDC 2007. McHugh has developed many AI and gameplay systems for DARKSPORE, including the ability system that will be discussed at AIIDE 2011. She will describe how the ability system enabled the team to easily make hundreds of cool and unique heroes and enemies. McHugh loves playing games, and especially likes online cooperative experiences.

Daniel Kline has been an AI and Game Programmer and Designer for consoles since 2001. He has shipped 6 titles and developed 10 titles, working with companies such as Activision, Electronic Arts, Blizzard, LucasArts, and Midway. He was Head AI and Gameplay Engineer on Star Wars: Force Unleashed, AI programmer on Diablo 3, and the engineer and designer of the first two levels of Call of Duty: Finest Hour. He has spent over 5 years doing Interactive Storytelling research, design, programming, and pre-production for 3 different companies and AAA blockbuster titles. He is currently working on Darkspore at Maxis as a Software Engineer.

This talk will discuss the techniques used to create and evaluate the NPC (non-player character) behavior in the popular Dead Space action-horror franchise. Scaring the players is an essential component of the Dead Space universe. In order to create scary situations the AI systems and even individual units must communicate state and events to each other as well as other systems. Gascoigne will present examples of how this information is used by the audio, level scripting, and lighting systems. He will also give an overview of the decision-making, sensing, knowledge representation, path-following, and debugging techniques used in the titles. The evaluation portion of the talk will cover the production methods used by the Dead Space 2 team to create new NPCs with a special emphasis on the use of focus testing. Specific examples like the Stalker enemy from Dead Space 2 will be examined in detail.

Louis Gascoigne is a game programmer at Visceral Games Redwood Shores, an Electronic Arts studio. He has 14 years of experience working on commercial games and is credited on the following titles at Electronic Arts: Dead Space 2, Dead Space, The Simpsons Game, The Godfather, From Russia With Love, Everything or Nothing, and Agent Under Fire. His primary interest is in making games fun to play by any means necessary and shipping them to (hopefully) large audiences.

The RTS genre has evolved considerably since 1993 when Warcraft: Orcs & Humans was first in production. "Evolution of RTS AI" will go back to the beginning and look at a few programming challenges that began with the Warcraft AI, and examine how the challenges and solutions evolved over the last 18 years, or in a few cases why things have not evolved at all. Bob Fitch was the primary AI programmer for Warcaft, Warcraft II, Wacraft III, Starcraft, Brood War, and Starcraft II. He will walk each topic through from Warcraft to Starcraft II while discussing each problem, what he wanted to do with each game, what he ended up doing, and why each decision was made. The topics that might be covered could include general strategy, attack wave management, base building, cheating, time slicing, pathing, transporting, difficulty levels, scouting, and maybe more.

Bob Fitch started programming games when he was 8 years old, working on an Atari 2600. He began making games professionally in 1992, programming Rock 'N' Roll Racing while working for Silicon & Synapse. That tiny company would later grow by more than two orders of magnitude and rename itself Blizzard Entertainment. Fitch designed or programmed part of every Blizzard release, including a Lead Programming role throughout the Starcraft franchise. In his free time, Fitch is the DM for a Dungeons & Dragons campaign, plays ice hockey, and continues his attempt to master Texas Hold'Em poker by hosting the Blizzard Home Game.