AI has become very impressive lately and continues to expand its capabilities. Let’s discuss how artificial intelligence might influence game development in the near future, from automated asset creation to personalised player experiences.
Artificial intelligence (AI) has begun rapidly transforming various industries (in addition to our private lives), and game development is no exception.
The integration of AI in game dev promises to bring about significant improvements in terms of efficiency, creativity and player experience. By automating complex tasks, providing deep personalisation and enhancing overall game quality, AI will very likely become an indispensable tool for developers (even though Nintendo has stated they will not make use of AI for, um, reasons? We’ll see how long that lasts).
9 Key Areas of AI Enhancements in the Gaming Industry
So, let’s see — where may game developers be using AI tools most successfully? And very soon? I think it could be here:
1) Asset Creation
Artificial intelligence could be used to (drastically!) speed up the creation of game assets automating at least parts of the process.
Traditionally, creating high-quality assets — such as character models, environments, textures or animations — has been a talent-requiring and time-consuming task. AI can accelerate these tasks in several ways:
- Automated Generation: With AI algorithms, artists can already generate a wide range of assets simply by writing some words and pressing Enter. Using AI models trained on vast amounts of data (photos, 3D models, icons and so on), anyone can create detailed textures, realistic environments and even characters with all of their animations.
- Style Transfer: AI can apply specific art styles to existing assets, transforming them to match the visual concept of the new game. For example, it should be possible to create a darker, more mature version of Game Master Plus within a few weeks using advanced AI models in the near future.
- Animations / Motion Capture: Up until now, creating realistic animations required motion capture in a studio, with sensors being attached to actors and cameras all around. AI-driven tools can simplify this to a great extent by generating animations from scratch (that is, from its training data) and/or from a simple video recorded with a smartphone.
- Asset Optimisation: By automatically adjusting polygon counts, texture resolutions and other parameters, AI can ensure that assets look good and are loaded quickly across a variety of devices.
- Quality Control: AI can assist in identifying issues such as inconsistent textures, clipping or incorrect lighting. By flagging (or even correcting) these issues early on, AI helps maintain a high standard of visual fidelity throughout the development process, which should be a novum for many studios. It probably helps in terms of motivation if your work-in-progress version of the game looks pretty awesome already.
The integration of AI in asset creation not only speeds up the development process but also allows smaller studios to produce high-quality games that can compete with larger, resource-rich companies.
Very soon, there will be no limitations in regards to the number and types of assets available to developers — including myself!
2) Avatars
Avatars aren’t that different from other assets, and so, providing a more personalised and immersive experience in terms of how the player and other characters look shouldn’t be a problem. Here’s how AI will most likely make an impact:
- Customisation Options: AI can offer a wide range of customisation options for characters by generating unique facial features, body types and clothing styles based on player preferences. This allows players to create avatars that closely match their imagination or even their own appearance. There will probably be some censorship, though, preventing you from creating Link and Zelda as the main protagonists for Final Fantasy XX, the American president as the arch villain or a nudist sidekick.
- Dynamic Changes and Realism: AI can enable avatars to adapt to the environment and the player’s actions. For example, avatars can change their appearance more accurately based on weather conditions or where exactly the blood of the enemies has splattered. Injuries of the player character may also alter the animations dynamically. The player’s aggressiveness or caution could serve as an input for the avatar’s expression.
I think games of today will age badly and will, ironically, feel very artificial once you’re accustomed to AI-generated and AI-updated character models.
3) Procedural Generation
Procedural generation is a technique where game content is created automatically, that is, algorithmically, during play rather than manually beforehand, leading to unique and often unpredictable outcomes. The use of AI can take procedural generation to a new level by creating more complex and varied content at a high quality. This will affect the following disciplines:
Level Design
Traditionally, level design has been a complicated process requiring a great deal of knowledge, creativity and time for fine-tuning. AI can streamline this process by generating levels that adhere to specific design or structural principles while introducing novel patterns and elements.
That way, AI could not only create an entire dungeon layout from a text prompt, but also the objects, textures, lighting and audio.
Story and Quests
Beyond the game environments, AI can also contribute to the narrative aspect and fill the automatically created game world with life and meaning.
Procedural generation of storylines and sidequests can lead to infinite replayability, with AI crafting unique scenarios based on player choices. This creates a narrative experience personalised to an extent never seen before.
Abilities, Items, Enemies
AI can also be used to generate new abilities, items and enemies. And by this I don’t mean adjusting parameters for each individual player. I mean, for example, …
- spells based on the mage’s elemental preferences and strategy in battle,
- an ultimate weapon designed and named in a way that reflects the hero’s dialogue choices or
- a group of enemies fit for a dungeon that is being procedurally generated.
***
As you can see, there are many possibilities — which not only enhance the player experience, but could also be fun to experiment with as a developer!
4) NPC Behaviour and Dialogue
Non-player characters are vital (haha) for creating immersive and engaging game worlds that feel alive. Guess what, AI can revolutionise NPCs as well!
Reactions
AI can generate realistic behaviour and emotions for non-player characters, making them way more lifelike.
By analysing player interactions and environmental cues, AI can adjust the avatar’s expressions, gestures and movements to put a more believable and engaging character into the game world.
Routines
AI could also create routines for daily activities and interactions with other NPCs. For instance, an NPC in a village might wake up, go to work, eat and drink and socialise with friends and family, all while reacting to the player’s presence or significant in-game events (like explosions! Or something).
A simple text prompt might be all you need as a developer, saving you hours, days or maybe weeks of tinkering with complex code.
Personality
AI can enable NPCs to adapt their behaviour based on the player’s actions and decisions. This means that NPCs can remember past interactions and change their attitude towards the player accordingly.
For example, if a player consistently helps a particular NPC, that character might become more friendly and offer exclusive quests or items. Conversely, hostile actions might result in NPCs becoming more cautious or even antagonistic.
Voice
Natural language processing (NLP) advancements and voice generation (see OpenAI’s GPT-4o Advanced Voice feature) allow NPCs to engage in more natural and fluid conversations with players: AI-driven dialogue systems could understand and generate human language, making interactions feel a lot more organic.
Players could also engage in conversations where NPCs understand context, have certain knowledge about the game world, remember previous dialogues and respond appropriately. This could lead to deeper storytelling, interesting puzzles (personalised riddles with personalised hints received from various NPCs, for instance) and, all in all and once again, a more immersive player experience. Just be careful not to get attached to any NPC!
5) Enemy and Ally Behaviour
AI can significantly enhance the complexity and realism of enemy and ally behaviour too, of course. Here are a few ideas what AI should prove useful for:
Adaptive Strategies and Tactics
Enemies controlled by true AI can learn from player actions and adapt their strategies accordingly. For example, if a player consistently uses a particular tactic to defeat enemies, AI can adjust enemy behaviour to counteract that tactic, requiring players to stay alert and continuously evolve their methods. The main advantage here is, once again, that developers don’t need to think of each scenario and write code for it.
AI can improve the coordination among enemy and ally groups. Enemies can work together more effectively, using flanking maneuvers, setting traps and providing cover in a shooter game, for example. Similarly, AI-controlled allies can better support the player, reacting to the player’s situation and assisting more intelligently. ATK-buffing the mage should be a thing of the past.
Dynamic Adjustment of Intelligence
AI can even dynamically adjust the sophistication at which enemies and allies operate. For example, in an Action RPG, a Confuse or Charm debuff may completely change an enemy’s or ally’s behaviour.
All devs need to do is describe what a confused or charmed character would be like. AI could also let enemies make believable mistakes on purpose in situations where the player is struggling too much.
Emotional Responses
AI can add emotional responses to in-game events. Allied characters or enemies might show fear when outnumbered or become enraged if their comrades are defeated, altering their behaviour in combat.
Allies could express relief when saved from danger, or frustration when failing to achieve objectives, adding even more depth to their personalities and interactions.
***
Many of these things are, in principle, possible today, but require a lot of scripting — the common theme here, as it seems. With true AI, on the other hand, we might see complex behaviour as outlined above even in the simplest games.
6) Dynamic Audio
AI can enhance game audio by creating more immersive and adaptive soundscapes. Here are some ideas what AI could do:
Procedural Music Generation
AI can generate music on the fly, creating soundtracks that adapt to the environment and situation.
For instance, the intensity of the music could increase during battle and shift to a more serene tone during exploration, using a different instrumentation upon entering a cave. This has been done before, of course, but now it’s just a text prompt for the developer. If that is even necessary.
Oh, and AI could adapt to the player’s musical preferences, of course, maybe adding some distorted guitars to the battle music or using electronic beats instead of an acoustic drum set.
Environmental Soundscapes
AI can create realistic environmental audio by dynamically generating sounds based on context. This includes ambient noises like wind, water and wildlife, weather conditions and more.
Sound Effects
AI can generate sound effects too. For example, the sound of footsteps should vary depending on the surface (wood, grass, metal) and the character’s speed. Similarly, sounds should change based on the environment, such as echoes in a cave.
Devs don’t need to find the right sound effects and add them to the project; instead, they just specify materials, and AI does the rest. (Maybe AI could also just recognise the materials by vision, making things even easier.)
Voice Acting
AI can already generate voices, so it could also generate voice lines for characters.
It could copy voices of well-known voice actors if they agree to it and are compensated for it.
My guess would be that studios will continue hiring voice actors for the major roles in a game while using AI to give every single NPC their own voice and fully voiced dialogue.
Oh, all the lines could be localised too! Carina Reeves / Fiora could, theoretically, speak German in another Xenoblade Chronicles remake one day, although I think the British accent constitutes a large part of the appeal.
***
By leveraging AI, game developers can create richer, more adaptive audio experiences that enhance the overall immersion and enjoyment of their games even more. Combine this with everything outlined above and you have almost got a new medium, wouldn’t you say?
It’s going to be crazy.
7) Fully personalised Experiences
In addition to personalised avatars and customisation in terms of story and music, there’s even more that can be done, AI has the potential to tailor games to the individual preferences and abilities of players in many other regards.
Let me just submit this short, but incohesive list of items I didn’t know where else to put:
Level Design
AI could create only levels that the player likes. For example, if a player enjoys solving puzzles, the game could automatically generate more puzzle-heavy dungeons. If the player dislikes underwater levels (as many do), there won’t be any in the game. The game could even model the starting town after the street the player lives in.
Abilities and Items
AI could create abilities or items that support a player’s preferred way of handling challenges. A player specialising in stealth, for example, might find more abilities and weapons that support this playstyle.
Difficulty
AI could adjust the difficulty dynamically to ensure the game is challenging but not frustrating at all times.
This might mean that enemies become smarter or more numerous if the player performs well, or that more healing items appear if the player is struggling.
Again, I know this is possible already, but it’s an area where AI can save developers a lot of trial and error.
***
With the introduction of AI-personalised experiences, games may become more like sandboxes: You don’t buy an interactive film anymore, but a game world where events will take place that are more or less guaranteed to be engaging for you, specifically.
The role of the developer will change from a director to an architect.
8) Playtesting, Bugfixing, Balancing
Playtesting, bugfixing and balancing are essential parts of game development that can be significantly accelerated by AI.
Here’s what it could do:
Immediate Bugfixing
AI can analyse code and identify potential errors or performance issues even before playtesting. Maybe it will be possible to fix bugs actual players encounter? The game would fix itself on the fly, so to speak, and the player could continue playing as if nothing had happened. No more crashes that result in progress lost!
Automated Playtesting
Why not let AI play through the game (even at a higher speed) and test various scenarios to ensure all game mechanics work as intended? This would a) reduce the number of human testers needed and b) discover bugs no human tester would have found.
Real-Time Balancing
Sure, AI can help balance the game by analysing data from playtests and identifying areas where the game might be too difficult or too easy.
But why not take it a step further?
AI could make adjustments while playtesting and save the new values right away! Combine this with AI playthroughs and you’ve got the perfect balancing within hours.
Of course, this is only necessary as long as games don’t use dynamic difficulty as outlined above.
***
Bugs, glitches and crashes should be a thing of the past once AI error detection has been integrated into the development process of just about any game.
But – will we not maybe miss them a bit? There have been hilarious bugs in the past as well as speedrunning methods based on glitches, after all.
Could we become nostalgic for bugs? Who knows. It might, indeed, be the case that some developers insert ’errors’ on purpose — as a new type of easter egg, if you will. A strange thought, right?
9) Game Analytics and Suggestions
In the near future, AI will help analyse even larger amounts of data than today. Devs will more clearly understand player behaviour, which allows for more informed design decisions. Here’s how AI-driven game analytics and AI-generated suggestions for improvement based upon the results will impact game development:
- Player Behaviour Analysis: AI can track in-game actions, player choices and progression patterns. By understanding these, developers can identify which parts of the game are engaging or frustrating, which abilities are used most often or which pieces of equipment are ignored, maybe because they lack interesting effects.
- Hints: AI could provide suggestions not only to the dev, but also to players, enhancing their gaming experience. For example, if a player is stuck fighting a boss, AI could recommend abilities or items that complement their playstyle and that were used successfully by like-minded others.
- Content Suggestions: Based on player preferences, AI could recommend additions to the game such as DLC, expansions or in-game events.
- Community Feedback Analysis: AI can process and analyse feedback from player reviews, forums and social media. By understanding common themes and sentiments, developers can prioritise updates and improvements that address the community’s most pressing concerns and desires.
- Monetisation Insights: This seems controversial, but AI can analyse spending habits and suggest optimal monetisation strategies. This includes identifying the most appealing in-game purchases, timing for offers and potential areas for microtransactions. Hooray?
- Player Retention: AI-driven analytics can help developers identify patterns that lead to players abandoning the game. By understanding these patterns, developers can implement changes to retain players, such as adjusting skills and items, introducing new challenges at exactly the right difficulty level or offering personalised rewards.
With AI, you don’t necessarily need data analysts and experts for integrating data sources. Any dev may be able to collect player data and extract some important insights from it.
Okay, what does all of that mean?
Let’s summarise.
The integration of AI into game development signifies a major shift in how games are created, played and optimised. Here’s a comprehensive look at the implications:
Efficiency
AI will dramatically enhance the efficiency of game development processes.
Automated asset creation, procedural generation and dynamic audio production, for example, will reduce the time and effort required to put out a game.
This efficiency allows developers to focus more on the creative aspects, arguably the most interesting part of development.
Immersion
With AI-driven NPC behaviour, dialogue and adaptive enemy tactics, game worlds will become more lifelike and thus immersive.
NPCs will have complex routines and show realistic emotions, creating believable interactions.
Enemies and allies will develop new strategies and implement new tactics on the fly — without lines and lines of code scripting behaviour for every conceivable scenario.
Personalisation
The ability to tailor game content to individual preferences will result in highly personalised gaming experiences.
From custom avatars and dynamic difficulty adjustments to original storylines and in-game recommendations, players will enjoy games that feel uniquely designed for them, enhancing engagement and satisfaction.
Continuous Improvement
AI-driven analytics will enable continuous improvement of games: By analysing player behaviour and feedback, developers can make data-driven adjustments to optimise gameplay elements, balance difficulty and introduce new content.
These ongoing changes ensure that games remain relevant and enjoyable over time.
Enhanced and Audio Quality
Optimised assets and dynamic audio will contribute to a richer sensory experience.
Procedurally generated music and environmental soundscapes will adapt to in-game events, creating an extremely immersive atmosphere (especially with headphones on).
High-quality assets optimised for various devices will ensure that games look great across different platforms.
Removal of Barriers
Smaller studios and indie developers will benefit significantly from AI tools, as they can produce high-quality games without the resources typically required. This democratisation of game development will lead to a more diverse landscape, with a wider range of ideas being represented — which, of course, also has its downsides as it increases competition.
Challenges and Considerations
While the benefits of AI in game development are substantial, the integration also poses challenges.
Ensuring an ethical use of AI, maintaining player privacy and addressing potential hazards and biases in AI algorithms will be crucial.
Additionally, the potential for over-reliance on AI might also stifle human creativity if not managed carefully, especially for larger, more risk-averse studios.
Conclusion
The integration of AI in game development is set to transform the industry by facilitating major changes in regards to a multitude of aspects – for developers themselves as well as for players.
In short: Developers will be able to focus on tasks they enjoy, players will experience more immersive and personalised games.
However, alongside these advancements come some challenges. In my view, it will be especially important to avoid over-reliance on AI, especially in AAA (or slightly smaller) studios.
But while there’s certainly the risk of ‘professional’ studios accidently streamlining their games due to AI improvements (or what looks like it), there will also be many ‘amateur’ game developers who complement their brilliant initial idea with knowledge about good game design stored in AI models as well as automatic adjustments or at least data-driven suggestions.
The future is bright and colourful.
***
Question: Are you looking forward to AI-enhanced games? As a dev? As a player? What are you looking forward to the most? What would be the downsides of AI in your view?