In this blog post, we’ll explore the current state and future direction of artificial intelligence in the gaming industry, as well as the future of game development.
What Role Does Artificial Intelligence Play in the Gaming Industry?
Since “AlphaGo” defeated Lee Sedol in 2016, interest in artificial intelligence has surged worldwide. Since then, AI has expanded its scope beyond Go to various other games, and its potential applications in real-time and online games are drawing particular attention. So, to what extent has AI advanced in online games—as opposed to board games like Go? What role can AI play in a gaming environment where countless variables exist?
AI in games is not limited to simply serving as an opponent for players. It also performs various roles that support game operations. It analyzes players’ skill levels to match them with appropriate opponents, maintains game balance, and detects cheating programs or abnormal gameplay. Representative examples include the ranked matchmaking system in “League of Legends” and the abnormal gameplay detection system in “MapleStory.”
‘League of Legends’’ ranked system provides a fair gaming environment by analyzing players’ skill levels and pairing users of similar skill to compete against one another.
Furthermore, various AI technologies are utilized to detect and penalize users who employ illegal programs, such as auto-hunting or auto-gathering tools.
How have NPCs and game AI evolved?
Currently, most games feature AI characters known as NPCs (Non-Player Characters). While there are traditional game AIs that compete against humans, such as in chess or checkers, games have recently emerged in which NPCs react in various ways depending on the player’s actions. Representative examples include “Black & White” and “The Sims.”
“Black & White” is a game centered on the player taking on the role of a god to gather followers and expand their territory, and the world and NPCs react differently depending on the player’s actions.
‘The Sims’ is a leading life simulation game series that simulates human daily life, designed so that characters behave appropriately in various situations.
Agents used in game AI are generally classified into reactive agents, deliberative agents, and learning agents based on their level of intelligence. Reactive agents are the most basic type, responding immediately to stimuli from their surroundings. Deliberative agents analyze input information, formulate plans, and then decide on their actions. Learning agents are characterized by their ability to continuously improve their performance based on past experiences.
In the past, most games utilized AI at the reactive or deliberative agent level. This is because games must handle a large number of NPCs simultaneously, and applying a higher level of AI would significantly increase computational and development costs. Furthermore, since learning agents require high-performance hardware, they were often used only in certain research projects or for limited content.
However, with recent significant improvements in AI computational performance and advancements in technologies such as reinforcement learning, large-scale language models, and generative AI, the scope of application for learning agents is gradually expanding.
How are learning-based AI and procedural content generation transforming games?
Unlike board games, where moves are made in turn-based sequences, online games involve countless situations unfolding in real time. Players and NPCs must perform various actions within a short timeframe, requiring more complex decision-making abilities. To address this environment, AI utilizing reinforcement learning has begun to be applied to games.
A prime example is ‘Blade & Soul’, which introduced AI based on learning agents into its “Tower of Infinity” content. This AI analyzes its performance through battles with players and adjusts the difficulty of subsequent opponents based on those results. Initially, the AI accumulated data by repeatedly battling against other AI instances, and later utilized data from actual players for training. Additionally, rewards and penalties were carefully designed to maintain user engagement while efficiently utilizing server resources.
This allowed the AI to continuously improve without placing a significant burden on the game’s performance. Rather than repeating predetermined attack patterns like traditional NPCs, the AI selects various strategies based on player behavior, allowing users to move beyond memorizing strategy guides and experience more suspenseful gameplay.
So far, we’ve looked at NPC-centric AI that assists players. However, “Procedural Content Generation (PCG)” technology—where AI generates game content itself—has also been advancing rapidly in recent years. This technology allows algorithms and AI to automatically generate maps, characters, items, quests, and more, rather than having developers create all content manually.
Since AI creates new content each time based on specific rules and algorithms, players can have a fresh experience every time they replay the same game. Prime examples include the random map generation system in ‘Spelunky’ and the Nemesis System in ‘Shadow of Mordor’.
‘Spelunky’ is a game where players explore caves and collect treasure; the layout of the map and the placement of obstacles change with every playthrough, resulting in unpredictable gameplay.
In ‘Shadow of Mordor’, the Nemesis System allows enemy characters to retain their own memories, ranks, and personalities, forming unique relationships with the player—such as evolving or seeking revenge based on their interactions. These systems provide a new gaming experience that differs from the traditional approach of simply facing the same enemies over and over.
Procedural content generation also offers significant advantages to developers. It reduces the burden of having to create all the vast content manually, thereby increasing development efficiency and extending the game’s lifespan. Additionally, players can experience new environments and situations each time, which helps reduce the boredom that comes with repeated play. Recently, with the advancement of generative AI technology, the scope of procedural content generation has expanded beyond just map creation to include various areas such as dialogue, quests, NPC behavior, and voice generation.
Why is artificial intelligence important for the future of the gaming industry?
Nevertheless, the development of artificial intelligence in the gaming industry has not spread as rapidly as expected. This is because developing AI requires significant time, cost, and specialized personnel. In particular, for large-scale games, securing data to train AI and conducting repetitive testing processes are essential, resulting in a considerable development burden.
Furthermore, it was often believed that AI in games only needed to provide the level of functionality necessary for players to enjoy the game. Consequently, some developers did not feel a strong need to actively pursue the development of advanced AI.
However, trends in the gaming industry are changing rapidly. Several global game developers are already building their own AI technologies and algorithms, and are applying AI to various areas—such as NPC dialogue, content creation, game testing, and operational automation—using generative AI, reinforcement learning, and large language models. These technologies are establishing themselves as key elements that not only improve development efficiency but also provide players with a more natural and immersive experience.
If these technologies become commercialized more reliably in the future, game companies that rely solely on traditional methods will likely struggle to maintain their competitiveness. Conversely, companies that actively research AI technologies and incorporate them into the game development process will be able to secure a stronger competitive edge in the evolving market.
Of course, not every game necessarily requires cutting-edge AI. However, as games grow in scale and user expectations rise, AI is becoming not just an option but a critical competitive factor. Companies that actively utilize AI from the development stage onward are better positioned to produce new content quickly, create more realistic gaming environments, and respond effectively to changing user demands.
Ultimately, in the gaming industry, AI is not merely a technology for making NPCs smarter. It is establishing itself as a core technology that enhances the efficiency of game development, provides players with new experiences, and determines the industry’s future competitiveness. As AI technology continues to advance, games are expected to evolve beyond simple entertainment into a new form of content that learns and grows alongside players.
