Once considered a niche sector, the computer and video game industry is reputedly worth US$30 billion today. As such, there is intense competition between gaming companies to produce games that beat their rivals to the post and capture a larger market share.
Artificial Intelligence, when used wisely, can significantly advance game play by making games more realistic. Half-Life, from Valve, was the first combat oriented game to incorporate a significant amount of Artificial Intelligence into its game play, therefore giving the gamers a far superior playing experience than its predecessors. Before Half-Life, combat games were typically one-dimensional. They hardly displayed an awareness of the player, let alone his style of play, consequently making the game too predictable and systematic.
Before the introduction of Artificial Intelligence into games, most AI opponents would just lurch forward, often right into the player's gunfire. However, the introduction of Artificial Intelligence techniques such as reinforcement learning, has made it possible for the computerized ‘enemies’ to duck behind pillars or around corners or when under gunfire, toss the player's grenades back, and even stand in for real players in multiplayer games.
Currently, many games have several levels of difficulty such as “Easy”, “Intermediate” and “Advanced” to cater to the varying abilities of different gamers. However, many game companies are currently involved in the making of games that can alter strategies and ploys to match a player's ability, instead of letting them choose a specific level of difficulty. Such games would not only be more dynamic, they would also be intelligent enough to uncover a player's winning tactics and hiding places, and take appropriate actions to thwart them!
Today, even the highest levels of most games follow a certain routine, which a seasoned gamer can easily uncover and exploit. However, with Artificial Intelligence, seasoned gamers can no longer be so smug, since they would constantly have to work out new strategies to win. This would prolong a game’s life considerably, and make the games more challenging, more fun and more complete.
The next wave of AI-based games will not only include dynamically changing enemies pitted against a player, but dynamically changing worlds and universes continuously changing to match the gamer’s level of ability. These AI controlled worlds will depend heavily on the decisions and actions a player makes, interacting with and reacting to him as he plays through. This will all but eliminate the disappointing, stereotypical plot lines found in most games. Some present day games are usually very linear—a leads to b: b leads to c: c leads to d—and the actions of the player have little bearing on the final outcome. In such games, each event and action must be scripted and coded. In this case, the gamer does not have much freedom because the programmer can’t possibly plan for everything.
With the new AI systems, only the basic interactions among characters and objects are planned out beforehand. Then, as the game observes and interacts with the gamer, the program is rewritten to include additional options and plot plans. The more the player plays, the more the program changes, providing an almost infinite number of twists and turns along the way. The ability to take into account consequences of previous actions will create far more realistic game scenarios and plot lines than ever before.
A classic game of Tic-tac-toe, with a little bit of 'intelligent' programming to prevent you from winning too often!