THE FUTURE OF AI IN GAME DEVELOPMENT

If you have ever played a video game, it is almost 100 percent likely that you have interacted with artificial intelligence (AI). Whether you are a fan of racing, strategy, shooters, or other genres, there will always be elements in games controlled by AI. Most often, however, it is associated with the behavior of characters, whether they are neutral merchants, enemies, or even animals.

So what is in-game artificial intelligence? We’ve dealt with how AI has been introduced into games, as well as the development of AI itself. In this article we will look at one of the most interesting and current technological trends — artificial intelligence. Get ready to dive into the world of AI and its capabilities!

What is AI?

Game artificial intelligence is a set of software techniques used in video games to create the illusion of intelligence in NPCs through character behavior. Game AI includes algorithms from control theory, robotics, computer graphics, and computer science in general.

AI is a technology that, through machine learning, allows a system to learn to analyze certain information in a virtual environment to produce behavior that is closer to human behavior. A few decades ago, something like this could only be found in science fiction, but now similar technologies are used everywhere.

AI in particular handles things in games such as:

- bot movement (rectilinear motion, motion with acceleration and turns, group motion, physics-aware motion, jumping, coordinated group motion, engine control in car simulators)

- pathfinding (Dijkstra’s algorithm, A* search algorithm, hierarchical pathfinding).

- decision making (decision tree, finite automata, fuzzy logic, Markov systems, goal-oriented behavior, rule-based systems, scripting)

- learning (quite a hot topic in AI, solved at the moment by parameter modification, action prediction, better decision making, better decision tree, reinforcement learning, and of course artificial neural networks) .

History of AI in games

The birth of artificial intelligence in video games began before the industry itself became an integral part of almost everyone’s life. One of the earliest and loudest precedents for the use of this technology in gaming dates back to the 1950s.

Then Alan Turing, known in the scientific community as the “father of computer science,” created the Turochamp software algorithm. The software was able to analyze the position of chess pieces on the board in order to choose the best possible move. The logic of the algorithm was based on only a few of the most basic chess rules, but it was only able to “think” two moves ahead. That is, Turochamp, of course, could play with real people, but certainly not beat professionals.

In the seventies, video games began to rapidly conquer the market, and all of their elements — to become better, and in exponential progression. And the developers paid special attention to the artificial intelligence from the very beginning of the game industry.

For example, AI controlled a racket in the game Pong, where it was already able to move based on the player’s actions. However, if we consider video games in general, the behavior of the enemies in them was quite primitive. Seventies microprocessors could not process a large number of sprites at the same time, so the game “lags”.

In the eighties, there was another well-known game — Pac-Man. This is the first video game, which introduced a system of route finding. With it, enemies could easily decipher the path that the player has chosen under certain circumstances. In addition, each of the ghosts in Pac-Man had unique patterns of behavior, which made confrontation with them more interesting.

And Donkey Kong from Nintendo gave players a whole bunch of different opponents with unique features and movements. The special pride of the developers was the smooth increase of the difficulty level, which pushed the gamers of that time to the limit of their abilities. However, artificial intelligence was still quite unreliable. It could neither learn from its mistakes, nor adapt to the players’ behavior, so the latter had to perfectly memorize specific patterns in order to move on — no indulgences were made by the authors.

By the end of the eighties the fascination with arcade machines began to wane. With the development of technology, personal computers and consoles began to actively appear in homes, which set a new direction for the development of video games. Games developed for these devices became more diverse because of the higher processor capabilities, but the AI still needed refinements. Developers had to take AI seriously, as it had become a prerequisite for a quality product.

Many new game genres emerged during this period. Among them, real-time strategies (RTS) have stimulated the development of in-game AI the most. In-game AI is extremely important for strategy games because it is the behavior of enemies in RTS that determines the level of interest, tension, and even excitement of gameplay. Thanks to strategy games, AI in games began to develop as an independent and important area of the gaming industry.

Although AI developers in the 1990s worked very hard to make NPCs look smart, these characters lacked one very important trait: the ability to learn. In most video games, NPC behavior patterns are pre-programmed; non-player characters are incapable of learning anything from players. The reason most NPCs don’t exhibit the ability to learn isn’t just because it’s hard to program. Often game creators prefer to avoid any unexpected NPC behavior that might somehow worsen the gamer’s gaming experience.

In the summer of 2017, it was revealed that Microsoft Research and Maluuba, a deep learning startup acquired by the corporation in early 2017, taught an artificial intelligence to play one of the most popular computer games of all time, Ms. Pac-Man. And not just taught it, they made it a champion, breaking the world record set by humans.

Developers of computer games use AI in varying degrees of elaboration. This forms the concept of “Game Artificial Intelligence”. The standard tasks of AI in games are to find a path in two- or three-dimensional space, simulate the behavior of a combat unit, calculate the right economic strategy, and so on.

What is AI for? How is it currently used?

AI has been around in the gaming industry for decades. But with the advent of modern tools and technologies such as graphics processing units (GPUs), advanced digital art software, and huge player data sets, the potential for both artificial intelligence and machine learning has increased dramatically!

Below are the main AI/ML implementations in video games:

1. Smarter NPCs.
Non-game characters (NPCs) are characters in the game other than the main player. Traditionally, NPCs were programmed with predetermined actions using a finite state machine. This meant that their actions were related to the storyline or in response to the player’s actions, so the NPCs’ actions were limited and predictable.

2. Dynamic Rendering.
One of the problems that video game companies are trying to eliminate with AI and machine learning is perspective distortion. This phenomenon occurs when an object looks good when the player is far away, but distorts and becomes pixelated when the player approaches said object. Game companies use machine learning algorithms to dynamically improve images and rendering. This will prevent the effect of image distortion and allow the object to look better when it is closer to the player.

3. Dialog Generation And Realistic Interactions
We’ve already seen how AI and ML can be used to improve NPC actions. However, these techniques can also be used to improve gameplay by formulating more accurate and realistic NPC responses.

4. World Generation.
Another powerful application of machine learning in game development is world generation. A number of popular games such as Minecraft and the Grand Theft Auto series use an open world game scenario.

5. Creating Immersive Games.

One of the top priorities for video game developers is to create a game that is as immersive and close to the real world as possible. However, simulating the real world can be an incredibly difficult process.

Challenges of current AI

U.S. Presidential Administration Releases 5 Provisions to Protect Humans from AI .

On Oct. 7, 2022, the White House Office of Science and Technology Policy (OSTP) released five regulations to guide the design, use and implementation of automated systems. The document comes as more voices join the call for action to protect people from the technology as artificial intelligence evolves. The danger, according to experts, is that neural networks easily become biased, unethical and dangerous.

1. Safe and efficient systems

The user must be protected from unsafe or inefficient systems. Automated systems should be developed in consultation with various communities, stakeholders, and subject matter experts to identify problems, risks, and potential impacts of the system. Systems should be tested prior to deployment to identify and mitigate risks and should be continuously monitored to demonstrate their safety and effectiveness.

2. Protection against algorithmic discrimination

The user must not face discrimination by algorithms, and systems must be used and designed on the principles of equality. Depending on the particular circumstances, algorithmic discrimination may violate legal protections. Designers, developers, and implementers of automated systems should take proactive and consistent steps to protect individuals and communities from algorithmic discrimination and to use and design systems on an equitable basis.

3. Data privacy

The user must be protected from data misuse through built-in safeguards, and the user must have the right to control how their data is used. Designers, developers, and implementers of automated systems should seek the user’s permission and respect their decisions regarding the collection, use, access, transfer, and deletion of their data in appropriate ways and to the maximum extent possible; if this is not possible, alternative design-based privacy protections should be used.

4. Notification and clarification

The user needs to know that the automated system is being used and understand how and why it contributes to the results that affect them. Designers, developers, and implementers of automated systems should provide publicly available documentation in plain language that includes a clear description of the overall functioning of the system and the role the automation plays, notice that such systems are used, the person or organization responsible for the system, and an explanation of the results, which should be clear, timely, and accessible.

5. Alternatives for the individual, decision making and fallback

The user should be able to opt out of services where appropriate and have access to a specialist who can quickly review and correct problems. The user should be able to opt out of automated systems in favor of a human alternative where appropriate.

Meta universes and NFT: the future of gaming

At some point the demands of game developers became largely satisfied by artificial intelligence, which we do not consider so intelligent today. The lack of big, noticeable leaps in the development of game AI is due to the fact that the underlying algorithms have not undergone radical changes.

Modern games still operate on the old fundamental concepts and methods in terms of AI, but use them on a large scale and with the advantage of the computing power of computers.

Most enemy AI actions can still be systematized and predicted by even the most unscrupulous gamer. Until recently, modern gaming artificial intelligence was only able to break its way to a sure victory in very narrow areas. For example, in chess.

Google’s DeepMind lab, Facebook’s AI research department* and other units around the world are hard at work teaching AI to play more complex video games. This includes everything from the Chinese board game Go and classic Atari projects to advanced cyber sports like Dota 2 and CS:GO.

They’re also trying to use neural networks to improve AI gaming, albeit as an experiment. There are some very famous recent examples, one of which was the AI that beat the professional Dota 2 team.

AI has several advantages over humans, such as the ability to multitask and react to things at lightning speed. Therefore, in some games AI developers have even had to deliberately underestimate its capabilities in order to improve gamers’ gaming experience.

Today’s developers don’t strive to create the most complex AI possible, but rather to use it successfully within game systems in order to achieve so-called emergent gameplay.

In the future, AI development in games will most likely not focus on creating more powerful NPCs so that they look for sophisticated ways to defeat players. Instead, development will focus on how to create a unique gaming experience for each gamer.

Gamers these days pay a lot of attention to detail — this includes not only the look and quality of the graphics, but also how vivid and interactive the game is in every way possible. And it’s the AI that can take the gaming experience to the next level. Maybe one day players won’t be able to tell if a character in the game is being controlled by artificial intelligence or another gamer.

In the world of meta universes, which Mark Zuckerberg and Epic Games are actively developing, everyone will need an avatar. Neural networks can be used to create such avatars. Now they already work in a similar way in metagames that have their own universes, like Roblox or Genshin Impact. In these games you can buy skins for weapons or clothes in special virtual stores.

There are several other trends in the development of the gaming industry:

- A rapidly growing market. According to Statista, as of 2021 the number of gamers worldwide exceeded 3 billion people, and this figure is constantly growing.

- Development of mobile games. Thanks to neural networks and 5G technology, mobile games will soon be as high-quality and popular as video games for PCs and consoles. For example, Genshin Impact, the top game of 2020 according to Google Play and Apple, works equally well on PCs and smartphones.

- The use of blockchain and NFT technology in games. One of the most popular NFT games is Waves Ducks, from Russia. In it, the user breeds ducks on a farm by purchasing EGG tokens, or “eggs,” to expand their farm. Players can buy and sell characters — the game has an NFT marketplace for this.

- Subscription-based monetization model. Apple, Microsoft, and Netflix, which recently added video games to its library, are already implementing it.

Let’s keep watching how AI-GameDev integration develops. We will see if our predictions turn out to be correct!