Not quite what you anticipated.
1997. IBM’s DeepBlue computer beats chess world champion Garry Kasparov.
2016. AlphaGo beats Lee Sedol, one of the best Go players in the world.
The scientific community responded enthusiastically to these advances in artificial intelligence in both 1997 and 2016. But why are there so many reasons to feel joyful when, in the end, all we’re talking about are board games?
Why, more generally, does this field’s study rely so heavily on finding solutions to videogame puzzles?
The obvious response is undoubtedly that videogames are challenging to solve because of this.
The difficulty is mostly caused by the fact that there are frequently multiple solutions and that they almost always compete against someone who will do anything to win.
Actually, when we give it some thought, what we do in every game is to adhere to a fundamental plan that, in our opinion, would enable us to achieve the ultimate triumph. This tactic frequently conflicts with the decisions the opposition makes, hence it is normal to alter one’s objectives throughout the course of a game.
As a result, even in games as straightforward as Tic Tac Toe, we would be unable to play without considering the opponent beyond our initial plan.
The complexity of the challenge is considerable due to the unpredictable nature of the situation and the difficulty of the game. In order to solve it, a machine must be given the autonomy necessary to compete with a person. even to defeat him, as was the case in the past.
We often learn how to handle issues in other professions by working through problems of this nature.
There is no doubt that games serve as a resource for AI research, but they also serve as a diversion for millions of gamers worldwide who, year after year, look for more and more sophisticated and avant-garde games.
The software companies that create these games have a responsibility to appease their fan bases and are constantly looking for new ways to employ technology to enhance the user experience.
The usage of AI in video games is currently quite limited, even though investments from this perspective are getting bigger—the 2013 release of GTA V cost 265 million dollars to produce and promote.
There must be a reason for this, of course.
AI and videogames today
The rationale for this is because the introduction of artificial intelligence into video games would cause a certain amount of unpredictability, running the danger of deviating from the plot’s intended course.
Additionally, an AI that can adjust to a specific player by, for example, watching their moves, remembering their errors, and picking up on their playing style, could render the game itself unplayable, giving the AI a significant advantage over other players and, as a result, destroying the primary goal of any videogame: the gaming experience.
For a game designer, it is crucial that the player has particular emotions at a certain moment in the game in order to make sense of the entire narrative that was created around that level. With artificial intelligence, there is a chance that this game’s turning point will occur sooner, later, or never at all, undoing all of the previous work.
Because of this, even if gamers like games that are unpredictable, taking this element too far may make it less intriguing and captivating. In fact, as we just saw, it might even damage the game itself.
So, what is generally done is to give only the illusion of intelligence. Make the gamers believe that they are facing opponents capable of reasoning, although these were actually designed to predict only the most common mistakes of the players.
In the end, with a little practice you can understand the opponent’s weaknesses and beat him in every game. This becomes possible because its weaknesses do not evolve and remain vulnerable to the same attacks.
The designers’ goal is to make this illusion invisible to players, which is precisely what they have accomplished. Consider this: an opponent who is too easy to defeat is foolish, whereas a perfect opponent can undoubtedly be attributed to a machine and hence to artificial intelligence.
When the opponent, who had been flawless up until that time, does something foolish, magic occurs, introducing a fallible mechanism that is common to humans.
As a result, the gamer sees it as realistic and continues to play without pausing to consider what is happening, sometimes even going so far as to defend the opponent’s poor decisions by pointing out how similar they are to those of an ordinary player.
With the introduction of the most recent uses of artificial intelligence, it is obvious that all of this would be unfeasible. simply because it would be monotonous if everything were flawless.
Because of this, we continue to favor designing a very complex system that offers endless game scenarios that could even happen to a small group of players over utilizing artificial intelligence to produce a lot more complicated but unpredictable result.
It’s ironic to consider that any player would like to play a game where freedom and unpredictability rule, but the software companies themselves travel in the opposite way since they understand that a game built by these standards would spell the end of the game.
It might be argued that this is an instance in which perfection is not required—in fact, it must be avoided because it is harmful.
Last but not least, games will continue to serve as a catalyst for artificial intelligence research for a very long time, offering progressively more difficult tasks that will result in beneficial discoveries in any sector.
Beyond research, it will be challenging to imagine a software company attempting to include the most recent AI applications into a videogame title in the near future because it is not the intended goal.