Introduction
Artificial Intelligence
Enter HyperReality
Brave New Worlds
Crafting Strategies
New Role-Playing Systems
2002 and Beyond
High-Tech Games
High-Tech Games: Pushing the Envelope in 2001 and Beyond
Artificial Intelligence

Black & White
Developer: Lionhead Studios
Publisher: Electronic Arts
Release date: Available now
Check latest prices
What's so high-tech about it: Its AI is the most complex of any computer game to date.

"The idea [for Black & White] was spawned by a Tamagotchi I was given as we were finishing Dungeon Keeper," says Peter Molyneux, the creator of Black & White. "I looked after it very carefully until one day it was drowned in a cup of coffee by Andy [Robson, who heads the test department at Lionhead]. I was quite upset, and then it dawned on me: If I could get that attached to an egg-shaped piece of plastic, how much more attached might I get to a computer creature that could respond to you and learn from you?"

screenshot
Click for full size image
Such was the inspiration for the latest game from Molyneux, the god game mastermind who launched the genre 12 years ago with Populous. Essentially, Black & White is a massive reworking of Populous, in which you traverse a fully rendered 3D landscape, and of gameplay principles found in Dungeon Keeper. But in Black & White, you teach your virtual-pet creature to do menial tasks, like watering villagers' crops or gathering materials to build structures.

The creature AI in Black & White is certainly more complex than that of Molyneux's Tamagotchi, and an early version of it was so sophisticated that it worked a little too well. Molyneux describes the magical yet odd moment when he and other members of Lionhead Studios witnessed the first creature in an early build of Black & White come alive on the computer monitor. The creature stood in its environment and then crouched its head down to bite its legs repeatedly. The programmers gradually realized that the newborn creature was trying to literally eat itself. "We had programmed him to eat the most nutritious substance around, but had omitted 'excluding yourself,'" Molyneux says.

screenshot
Click for full size image
Richard Evans, head of AI programming at Lionhead, explained how the impressive learning behavior of the creatures in Black & White works: A creature takes an inventory of locations in the game and the attributes of particular objects. From this, it creates "opinions" about what sorts of objects are most suitable for satisfying its various "desires," which in turn are based on mitigating factors. For instance, a creature's hunger desire starts to increase if it is low on energy, becomes bored, sees something tasty, or sees another creature eating.

"These different sources each have their own separate weighting, so you can train a creature to get hungry only when his energy is low or alternatively encourage him to eat whenever he is bored," says Evans. The creature then decides which desire is the most important to meet and which object is the most useful around to satisfy this. "To speed up the planning, the [creature's] search space is abstracted into two layers: large objects (towns, flocks of animals, citadels, [other] creatures) and small objects (villagers, rocks, [individual] animals). To keep things fast for a real-time game, [this] is staggered over a number of game turns. Once a plan has been decided on, it is broken down into atomic subactions."

Thus, the path that you take in the game's story, the way you train your creature (humanely or abusively), the reactions of the villagers toward it, and ambient sound and music all have an influence on how your creature's intellect grows--sometimes in unexpected ways. Molyneux says that during the game's development, one of his creatures enjoyed playing games of catch with another, tossing rocks back and forth between them. After a while, it appeared that the other creature became tired of playing. So when Molyneux's creature wasn't looking, it placed a rock in a fire to make it hot and then kicked it with its feet onto Molyneux's creature's personal rock stash. "My creature picked up the hot rock," he says, "and badly burnt its hands."

screenshot
Click for full size image
"The most important thing about the creature learning is that it represents a departure from the 'carrot and stick' approach," says Evans. "Let me explain: Suppose you want to teach your creature to attack only enemy villagers. Your critter arrives in a friendly village [and] behaves aggressively; you punish him severely. The trouble is that your creature has learned too much from the lesson, and now his aggressive urge has been permanently muted. To solve this problem, I introduced another much more important type of teaching: teaching by showing. [Your creature] sees the action you are performing, makes an intelligent guess at what your goal was, constructs a belief about the object you were acting on, and constructs a training episode from your action. The creature keeps a model of what goals the player has been trying to achieve. By using training by showing, you can always get your creature out of any [behavioral] rut he might be in."

In the end, the creatures' AI and interesting quirks were the direct result of how Molyneux likes to develop any game: by designing the technology needed for it at the same time as the rest of the game is being made. He prefers this approach in order to encourage every person on his team to be as creative as possible without being limited by preestablished design boundaries. "This style of development can take a bit longer," admits Molyneux, but the delay in Black & White has meant that "no one at Lionhead has ever said, 'that's impossible.'"
 

« Previous Page Next: Beyond real lies the hyperreal »