The term AI has been used in gaming for a long time, generally referring to behaviors and actions controlled by a computer within the game. Depending on the type of game, this could be non-player characters (NPCs), opponents, or other in-game elements. Essentially, it is when the computer attempts to play with intelligence, like another human player. Early gaming AIs were very basic and certainly nothing approaching artificial intelligence (AI) algorithms available today. Advanced AI and machine learning are increasingly being used in several gaming areas, however.
As well as the way in-game elements behave, there are AI and machine learning applications in gaming that you might not expect. Many people enjoy playing real money slots at online casinos, for example. It’s fun and gives the chance of big wins but you also expect it to be secure and to be played on a level playing field. AI can help in this by playing a part in the extensive cybersecurity used by online casino platforms. They can also be used to spot players deploying their own AI to play the system and win deceitfully.
Elsewhere in the gaming sphere, these technologies have several other uses. Voice recording and dialogue is one area of game development that can be costly and time-consuming. Blockbuster titles will still use well-known voice actors but AI-powered Text-to-Speech (TTS) and synthetic voices can help create stock dialogue, while automated translation systems can help localize games for other markets.
Motion tracking and haptic sensors (that produce lifelike sensations for the gamer) are becoming increasingly important as gaming moves further into the realms of virtual reality (VR) and augmented reality (AR). Computer vision algorithms are even making the graphics in high-spec games more realistic, letting you zoom in closer without losing detail.
And, of course, classic AI – the ability of a computer or artificial system to think like and often outplay a human – has come on in leaps and bounds. Chess was seen as an early benchmark of AI ability, but IBM’s supercomputer Deep Blue was able to beat the then reigning world champion as far back as 1997. Since then, AIs specializing in chess have far surpassed human abilities.
One of the current leading chess AIs is Stockfish and it has a rating that translates to a skill level around ten times greater than that of the best human player in the world. AIs have also beaten top human players in the Chinese strategy game of Go (which has more potential variations than chess) and have challenged eSport champions in hugely complex real-time strategy games like StarCraft II.
While the average gamer might not be directly benefitting from these high-profile challenges right now, the underlying research is being used to observe and improve the way AIs function. This could be used to create in-game AIs who genuinely react like their characters, adding far more realism and complexity to video games of the future.