We’ve been talking about AI in game development for a couple of articles already, mostly focusing on what artificial intelligence can do to make games more engaging. This time, we’d like to switch for a second to the developers’ point of view and explore the use of AI to make the job of creating games — mobile games specifically — easier, faster, and cheaper.
The competition on the video game market is fierce, more so if we’re talking about mobile games. Due to the high penetration rate of modern mobile devices and their ever-evolving tech characteristics, the video game market has been moving more and more to mobile-first and mobile-exclusive games.
There are a lot of perks mobile games offer to game creators compared to PC and console ones:
- High user penetration
- An opportunity to create never-ending games
- Longer user lifetime
- More frequent gaming sessions
- An opportunity for continuous revenue via ads and subscriptions
- Diversity of genres, some of which just won’t work outside mobile (like idle games, for example)
- Lower creation cost
- Shorter time-to-market
With the addition of cloud technology, the restriction put on developers by mobile devices’ tech specs has also been loosened somewhat; now, we see mobile games with superb graphics, open worlds, and MMO capabilities.
All this advancement created opportunities for smaller companies to enter the market and compete for user attention with the giants.
And using AI in game development can make it easier.
The role of AI in mobile game development
The AI disrupting the software development market is, today, an undeniable reality. And video game development is not an exception. Rather, it’s at the forefront of this particular trend.
We’ve covered this to some extent in our previous article, on how AI can enhance video games. We’ll repeat some of that information here. Some things AI can help game developers with include:
More realistic NPCs
By analyzing extensive data on human behavior, the AI technology can help game designers create much more diverse NPCs. And if you add machine learning to the mix, the NPCs can also be made to evolve as the game progresses and with regard to the player’s choices.
More detailed graphics
Upscaling, a technique that uses AI to enhance game art by using high-resolution photos, allows game artists to create stunning zoomable landscapes and lifelike textures.
Diversification of content
Generative AI services like ChatGPT enable game creators to “cover more ground” in less time. The job of creating a lot of various quests and events can be (and usually is) done by humans, but AI can facilitate the process.
The current quality of instant translation services is, at least in part, due to the use of machine learning and AI. Trained on human input, algorithms learn to translate with nuance. This creates a possibility for game developers to:
Additionally, AI-based speech generation allows small game development companies to save on voice actors for multiple languages.
Game research and ideation
One of the opportunities of AI in game design is the help in market research and ideation. By analyzing massive amounts of data on the internet, AI algorithms can help game producers understand the demand and make data-based decisions to make their games successful.
Testing and bug fixes
Early detection of bugs is critical on the market as competitive as mobile gaming. AI saves game developers a lot of headache, time, and money by quickly detecting not just existing but potential bugs.
By employing artificial intelligence in game design when you set up your monetization strategy, you can create personalized marketing offers for players. Such offers will, by virtue of AI and ML, adjust to each player’s style and preference, creating offers that are hard to decline.
Where traditional analytics already offer an abundance of data to enhance games, AI tools for analytics can help you make custom changes to cater to each player as well as to the player base as a whole.
Top AI tools for mobile game development
Above is just the tip of the iceberg that is the use of AI in game development. To help you understand a little better the multitude of application options, in this section, we will offer some examples. The market already has functioning services that help game developers do their work faster and more efficiently. We will be separating the services by application.
Game idea & research
If you want your idea to find its niche and achieve success, the first stage of game development is idea validation and research. In essence, no idea is bad or doomed from the start, but to bring its light out, you need to know:
- The market situation
- Your potential target audience
- Their expectations
- What’s trending and why
Armed with ample information and with the help of an experienced team, you’ll be able to tweak the idea so that it hits the market demand. And artificial intelligence makes this easier to do.
The primary function of AI across tech industries is search and analysis, that’s how AI tools are used in healthcare, for example. That’s also how generative AI “creates” things — it searches its training data for information related to the prompt, analyzes it, and puts together a mash-up. This function is extremely useful for game idea research.
Services like Ludo.ai employ intelligent algorithms to quickly gather information on the game market trends — past and current ones — and make predictions for the nearest future. While this job can be done by people, an algorithm will do it faster and at a bigger scale. Statistics, charts, trend patterns, popular game mechanics and elements — all this an AI can provide within seconds. Then you can go over the data, already narrowed down to the most essential bits, and make better decisions for your game’s success.
Moreover, such research can also be of help when you’re stuck, the inspiration isn’t there, and the idea isn’t fully formed in your head. When something is missing. Research algorithms can catch data from sources you might have missed in the pile — smaller ones, quieter ones. This way, AI can help fill the gaps, making your game idea fuller faster.
Services to use:
Game design and mechanics
During the development, your game designers will make assumptions about players’ actions. Such decisions are based on the designer’s experience and the data they gather through research. It’s also tested — that’s what A/B tests are for, you can offer two (or more) different test builds to a test group of users and see which one gets the best response. The “response” can vary:
- Length of a playing session
- Number of mistakes and why they happened
- The point of bounce (when players abandon the game)
And more. With the help of AI used in game development, it’s possible to predict players’ behavior and test without involving real people. Which mechanic should be used at each point? Where to pop-up paid offers? How will users react to each new feature you implement? Which event will be more suitable for the game and at what time to launch it?
Answering questions like this usually takes time, since you need to observe the target groups for some time. AI simulation, however, can make it faster.
One of the services gaining traction in this field lately is Machinations.io — a platform that offers a design-and-play predictions interface for game designers. The platform claims that it helps game developers ship 34% faster, increase KPIs, and avoid potential problems by predicting them.
Another one of the uses for generative AI in this stage of game development is that it can help game designers build and fine-tune game balance, as well as generate ideas for new levels and for player progression. Overall, making core and meta gameplay deeper and more entertaining is easier with the help of AI-backed tools.
Services to use:
This is the realm of generative AI. NPCs are the most uncreative creations in small- and medium-scale games. Making them truly interesting and diverse would take a lot of time for game designers and writers, and that is not sensible if the player encounters an NPC rarely or their interactions aren’t important for the full experience.
Companies that build AAA games spend years on each game, and months out of that on NPCs, to make them fleshed out. Smaller game companies can’t afford such timeframes. However, they now can create lifelike NPCs with personalities and goals fast — with the help of services like Inworld Character Engine and Charisma.ai.
Smaller players in the gaming industry can now offer unique experiences to their players without breaking a bank.
Services to use:
Game narrative & localization
Another use for generative AI and large language models (LLMs). The big name here would obviously be ChatGPT, but it’s not without competitors — the language models from Stability.ai are keeping the competition heated.
AI-powered LLMs can spin a story. We’ve mentioned several mobile games fully built on ChatGPT in our previous article, they include relationship simulators, story-generating adventures, and character creation games.
Now, there’s a big “but” here, an elephant in the room: The generative AI we have today is inconsistent and forgetful, so making up a story of any considerable length without self-repeating or self-contradicting is a challenge for LLMs.
Meaning, they need human supervision as a tool for game narrative writing. However, you can use AI in video game development to write short narratives (e.g., for events) and to assist your writers and speed up their work.
The second use for LLMs in game making is localization — instant translation into other languages. If you want your game to spread globally, it’s essential to have its interface and content in as many languages as you can.
Generative AI based on large language models, when properly trained, can do a better job than traditional neural machine translation (NMT) services. The reason being that localization of literary text is nuanced and often reliant on context. AI LLMs can be trained to recognize that.
Services to use:
Compared to other applications in game development, the number of art and asset related services with AI is huge (we’ve found 18 services by just scratching the surface). That is most likely due to the popularity of Midjourney and the like, as well as people’s general fascination with art and artists.
One of the tasks where game artists can get the most help from AI is the concepting stage. Visualizing characters and items often takes many tries to get right, and drawing anew each time is a hassle. Instead, artists can do it via text-to-image generation. It won’t be the end product, but it can help go through approval for visual style faster, which will allow artists to start on making the real thing sooner. (With which AI can also help.)
Artsy generative AI is a controversial topic, including in generating assets for games. However, the issue is already being addressed, and some services — like Scenario and Layer.ai, for example — already offer game asset creation by AI trained on your hired artists’ works. This means that you’re using the tool with explicit permission from the artist.
In cases like this, the AI algorithm is a huge help in completing your game faster: your artists can focus on the most important and intricate work while the algorithm creates assets for NPCs, equipment, background buildings, etc.
Another use of generative AI in game art is upscaling — the process of making the art more detailed by using high-resolution images to add textures and context-adherent details. Popular services in this niche include Stable Diffusion and Leonardo.ai. OpenAI’s DALL-E also has similar features.
Making game assets faster without compromising their quality and adherence to the same style would allow indie and small game development companies to release new games faster. And that helps them to survive the awfully competitive market.
Services to use:
- Layer AI
- Promethean AI
- Stable Diffusion
- Adobe Firefly
- Skybox AI
- DALL·E 2
- Texture Lab
Game sound and voices
Sound is an integral part of gaming, it can sometimes make or break the game. So it comes as no surprise to us that alongside generative AI in game making for art and text, there are also services that can create sound.
Even if you don’t need voiced dialogs, and even if your game is a small in scale idle title, it still requires multiple sounds to be entertaining:
- Background music
- Sounds for pop-ups
- Sounds to go with different actions
That’s where services like Splice, Meta’s AudioCraft, and Stability.ai’s Stable Audio come in. These services can generate music and sound effects from text prompts.
If your game is narrative- or character-driven, you’ll also need voice-overs. The best option would of course be to hire voice actors, but that’s an expensive endeavor, especially if you require it for multiple languages.
Platforms like Resemble.ai and Murf.ai offer text-to-speech generation with the help of AI and ML. You can make your characters sound distinct while staying within budget.
Services to use:
Game development & testing
You’ve probably heard that artificial intelligence in game development can help not just creatives, but coders too. But how? Well, algorithms can:
- Speed up code writing by using — yes — generative AI trained on code;
- Automate code writing for repetitive parts;
- Test code as it’s being written — by using bots and analytics tools;
- Analyze code for mistakes, inconsistencies, and integrity violation;
- Help onboard and upskill your developers.
Word of caution: As all other generative AI tools, code writing services aren’t perfect and require human supervision and correction. However, they’re already able to be of huge help to developers and quality assurance specialists, shortening games’ time-to-market and allowing re-allocation of effort for better results.
The most popular AI tools for coders and testers are:
- Unity’s own AI platform
- GitHub’s Copilot
- Mutable AI
- Towards GameDevGPT
- ChatGPT (mainly for code validation)
Best practices for using AI in game development
While it can already do a lot of things, the artificial intelligence as a technology is still in its budding stage, a lot of trial and error is involved when integrating it into your processes. Following are the advice we as a mobile game development company can give based on our monitoring of and participation in the AI integration processes.
Combine AI with humans
Whether algorithms will be able to replace humans in software and game development is still up to debate. Our own observation so far: Probably never, but definitely not any time soon.
Algorithms lack real creativity, can’t properly retain information on submitted output, and are generally inconsistent. In an industry as creative as game development, replacing humans with AI will not bring good results.
Hence, you need to work out a system where AI is a tool your game designers, artists, writers, developers, and testers wield as a sword and a shield. Do not send robots alone to the battlefield that is the gaming market if you want to be on the winning side.
Integrate in iterations
Jumping into the use of highly complex algorithms from the get-go will place a strain on your developers and game designers, and instead of assisting, AI will hinder your game creation.
Artificial intelligence is a complex technology, so when you implement it, start small and expand as you go. Your AI game developers will thank you for that, and your games will have higher success chances if you ease your team into new processes step by step.
Focus on adaptability
With all the newness of AI, it’s important to make your systems adaptable, since you never know how your team and your players will react to new workflows and AI-based features in your games.
In order to not lose time and money when something goes wrong (and something always does with things that are totally new), your systems need to be flexible and adaptable.
Use data religiously
Data is the king and the queen of commercial game development, more so when we’re talking about AI in game making. Analyze your team’s performance and your players’ behavior as you add AI into games and game creation processes, and use that data to adjust everything.
AI is evolving fast right now, and we’re yet to see the plateauing of this evolution process. Therefore, to employ it for the best results, if you’ve decided to use AI, you need to follow its evolution closely. If you don’t, you might miss on not just on new features and opportunities but on possible solutions to existing issues. And that might cost you dearly in the end.
How we can help in AI game development
Artificial intelligence isn’t one of the fleeting trends that come and go, forgotten after a year or two. What we’ve seen so far shows that it’s a huge leap in technology, similar to industrial and scientific revolutions that happen once or twice a century. It’s here to stay and evolve, making human lives easier in multiple ways.
We’re ready to board this train.
Right now, we at Mind Studios Games are in the middle of adopting AI in game development for our projects. We’re researching new services and trying new ways to make games more entertaining for our players. Our latest released game that uses artificial intelligence is the Beetle Riders 3D — an IO racing game where players can compete with friends or with AI-created opponents.
We’ve also been trying a bunch of AI services for internal processes and keeping an eye out for emerging trends in the industry. We’re testing the tools we mentioned above, and we’re actively using the ones we find the most suitable for each stage of our game development projects.
Being a small team of passionate game developers, we believe AI can be of huge help for anyone who wants to get out there into the game market but doesn’t have an AAA+ budget and a huge team.
If you’re looking for a team who’s already dabbled in AI technology, know their way around trends, and can keep your game development within budget while delivering a quality product — get in touch with us for a free consultation.