Zephyrnet Logo

The past, present and future of AI in gaming

Date:

Artificial Intelligence is surely one of the greatest buzzwords – or buzz phrases – of modern times. It inspires as much trepidation as excitement as we try to work out whether it will turn out to be a boon or a threat to our existence. After all, the idea of the machines taking over has long been fear and, today, we seem to be reaching a tipping point where it may become a realistic question to address.

Away from these doom-laden predictions that we may be architects of our own destruction, there’s no denying the fact that the use of Artificial Intelligence is proving to be an invaluable asset. This is particularly true of the gaming world in which it’s starting to play an increasingly important role.

But, while one might assume that AI in gaming is a relatively recent arrival, it has a longer history than many might imagine.

Early days

As long ago as 1952 programmers were experimenting with AI in a game called Nim, named after the computer, called Nimrod on which it was played. The game itself was like a virtual version of Jenga in which players had to remove matchsticks, one by one. It was undoubtedly basic, but it was here that the foundations for future developments in AI were laid.

Chess is a game that has long been beloved by mathematicians and computer scientists. So, it was not much of a surprise that one of the first high profile examples of the use of AI came in the 1990s with the chess-playing computer called Deep Blue.

It was set to challenge the world chess champion, Gary Kasparov, in 1996 over six games with the human player eventually winning 4-2. But Deep Blue became better at playing as time went on and won the rematch the next year 31/2-21/2. The fact that the computer had become better at chess without any physical human intervention was a source of great excitement.

More recently, a similar experiment was carried out by pitting a poker-playing computer against a number of professionals in a tournament format. Even though poker rules are very different from chess and there is also the need for more intuitive play, this computer also prevailed, again growing stronger and more proficient the more games that it played.

Making the random more random

Looking more specifically at video games themselves, the most common use of AI today is its role in controlling the Non-Playable Characters in games and it has started to allow them to behave in a far more random way than ever before. Using a variety of algorithms, many of which are based on one from the 1990s called the Finite State Machine, these allow for a decision tree to be used which then dictates the action.

The increasing sophistication of the algorithms being used today allows for many more options to be presented and a greater sense of randomness to be apparent in the game. One could argue that this just gives the impression of artificial intelligence being at work, however there is one area in which it is increasingly playing a part.

Because the whole purpose of a game is to be enjoyable and involving playing, AI is now being used to sense the ability of the player in question and to tailor certain games to suit. This increased “playability” is seen by many as the essential ingredient that will see players persist with a game far longer than a rigid one that doesn’t adapt in any way.

However, it’s a fine balance that needs to be struck to avoid games becoming too challenging or too simple to play. It’s perhaps this that is the biggest headache for games developers intending to introduce AI into the programming of a game.

Creating the narrative

One particularly exciting area that is currently being explored is in the field of RPG games like Dungeons and Dragons. The almost infinite nature of a game like this is the perfect format in which to explore whether AI can be used to work effectively in a collaborative, rather than a combative way.

This certainly involves taking on certain human characteristics as well as showing a level of empathic behaviour – tricky even for some people to achieve, let alone a machine. It also requires the ability to both set and follow a narrative and once a machine starts to do that it is on the verge of becoming an autonomous, creative entity.

AI-designed games

This vision of the future is taking shape even now thanks to an AI project called Angelina. This is the brainchild of a developer called Michael Cook who started working on it in 2011. By feeding a limited amount of information into the program, it can be left to its own devices to create a game across a number of genres from puzzles to adventures.

Other game companies like Ubisoft are taking it even further by experimenting how to integrate AI into producing triple-A video game design and working in collaboration with the University of Lyon.

It will obviously be quite a while until AI can be relied upon to create games that are as intricate, involved and nuanced as those designed by humans. But there’s no doubt that we are edging ever closer to that day.

So, before long, the question may not be whether a computer can beat a person, but the other way around.

spot_img

Latest Intelligence

spot_img