There was a time, a long time ago, when a newly born science wooed every single computer scientist and some enthusiasts. Artificial Intelligence(AI), which was born on 1956 grew in many areas only to face a very hard time when finding funding for any project seemed almost impossible. Yet years later it grew up again and turned into one of the main fields of study in the research scene.
Years after that, a new industry for entertainment was born and quickly transformed into something of its own. The video game industry started out with relatively simple games to entertain people with their gameplay. But soon it would even be able to make more than one jaw drop with graphics that every day approach more and more reality.
This industry found its very core at computer technologies, since it all reduced to a program running on a machine. Nevertheless, entire worlds were being created were virtual characters lived. And these characters, along with the world needed to behave, to act, to think. That was how slowly, an entertainment industry started to look for places where it could borrow ideas to solve its own problems.
From this point, AI started pouring into games to let characters plan on a RTS, find ways through unknown terrains and solve a wide variety of problems. One would think that this merging of areas would increase in time and that "rea AIl" would be both in the academic field and the video game industry. Yet over time the tendence has been the opposite.
Nowadays, games barely incorporate AI, being pathplannig apparently the be all end all of AI in games. And it is not even that good. Don't believe me?
Years after that, a new industry for entertainment was born and quickly transformed into something of its own. The video game industry started out with relatively simple games to entertain people with their gameplay. But soon it would even be able to make more than one jaw drop with graphics that every day approach more and more reality.
This industry found its very core at computer technologies, since it all reduced to a program running on a machine. Nevertheless, entire worlds were being created were virtual characters lived. And these characters, along with the world needed to behave, to act, to think. That was how slowly, an entertainment industry started to look for places where it could borrow ideas to solve its own problems.
From this point, AI started pouring into games to let characters plan on a RTS, find ways through unknown terrains and solve a wide variety of problems. One would think that this merging of areas would increase in time and that "rea AIl" would be both in the academic field and the video game industry. Yet over time the tendence has been the opposite.
Nowadays, games barely incorporate AI, being pathplannig apparently the be all end all of AI in games. And it is not even that good. Don't believe me?
We are now in a new era of gaming, where graphics have surpased the expectations of most, where novels and Hollywood have teached games to evolve into something more. And eventhough in this regard they have turned in the succesor of movies (novels-> movies -> games, but more of this on the next entry), they have relinquished their roots from computer science.
You may think that its not true at all, but when you love games and live in the academic field of AI, the full picture is quite evident. Bear with me a little longer to explain. Although AI looks more and more complex as time goes by, the reality is that the trees are growing. With this, i'm talking about decision trees:
You may think that its not true at all, but when you love games and live in the academic field of AI, the full picture is quite evident. Bear with me a little longer to explain. Although AI looks more and more complex as time goes by, the reality is that the trees are growing. With this, i'm talking about decision trees:
The one on the left is a simplified decision tree. It consists of several decisions along it that lead to an end output. If we pause a little to think about modern games that give you choices that affect the game, these are just big decision trees that move from one point to another. A big difference in how, for example, we humans are.
We are a very special kind, we can face two times the same situation and react different on each with actually a coherent argument on why we did that. But we never run on decision trees! The "tree" on the right of the image shows a little more how we acually make our decisions, and that is not even considering probabilities.
We are a very special kind, we can face two times the same situation and react different on each with actually a coherent argument on why we did that. But we never run on decision trees! The "tree" on the right of the image shows a little more how we acually make our decisions, and that is not even considering probabilities.
We as humans, change our mind, and change the minds of others! Change what surrounds us, and that affects others as well. Not everything is as simple as a binary decision, yes or no, good or bad, etc. AI in games have put more emphasis on these simple approaches, that is not something bad, but technology has advanced so much to still be using these archaic methods.
And for other tasks? Well cheating AI is pretty common in games. Enemies that go through walls when no one is watching, cars that stop actually racing and just get closer to the player when they are far away, etc. The list goes on and on. But by this point you might be wandering why I rant about this so much. And no, it is not to promote what we do at JARA, it's actually the other way around.
Why not take AI as it is (i mean the academic field) and apply it to video games? Imagine what could be done: games that change every time you hit new game, societies of NPC's that actually impact the world: a city that could grow to an empire and even change the story itself, story's that get written on the fly for each player, worlds that evolve, enemies that actually think and team up to trap and defeat you, enemies that adapt to your tactics and force you to find different approaches, characters that exhibit human emotions!
The list can go on forever, just imagine the unimaginable and thats what can be done, where we want to go. Or, being too naïve, where we want gaming to go.
Feel free to leave your comments and thoughts and have a very nice day.
Lex
And for other tasks? Well cheating AI is pretty common in games. Enemies that go through walls when no one is watching, cars that stop actually racing and just get closer to the player when they are far away, etc. The list goes on and on. But by this point you might be wandering why I rant about this so much. And no, it is not to promote what we do at JARA, it's actually the other way around.
Why not take AI as it is (i mean the academic field) and apply it to video games? Imagine what could be done: games that change every time you hit new game, societies of NPC's that actually impact the world: a city that could grow to an empire and even change the story itself, story's that get written on the fly for each player, worlds that evolve, enemies that actually think and team up to trap and defeat you, enemies that adapt to your tactics and force you to find different approaches, characters that exhibit human emotions!
The list can go on forever, just imagine the unimaginable and thats what can be done, where we want to go. Or, being too naïve, where we want gaming to go.
Feel free to leave your comments and thoughts and have a very nice day.
Lex
RSS Feed