Get relevant information on mobile marketing delivered to your inbox.
Back to blog

AI in Gaming: The Future of Player Engagement

Momchil Kyurkchiev Chief Strategy Officer & Chief Revenue Officer (West)
AI in Gaming: The Future of Player Engagement

These days, you can’t walk down the street without hearing about AI. It’s the latest buzzword but for good reason. It’s taking several industries by storm and gaming is no exception. Any function across the gaming production lifecycle can benefit from AI from art, design, and even live operations. Even though we’re in the very early stages of the impact of AI in the world of gaming, it has great promise. Let’s dig in further!

MIT defines artificial intelligence as “the ability for computers to imitate cognitive human functions such as learning and problem-solving.” In academia, there’s the famous ‘Turing Test,’ named after Alan Turing, who developed the test, where a human is asked to tell the difference between a human and a computer with a set of questions, and if the human cannot tell the difference, it passes the test. These days depending who you talk to, we are already there: several processing models have passed the test. But let’s talk more specifically about AI’s application in gaming and engagement.

In the gaming industry, there have been significant advancements in AI across various areas of the game production cycle, from art and design to coding and engagement.

Art and in particular generative art is perhaps the most prevalent AI example seen in popular media. In some ways, it’s an evolution of photoshopping a picture or using advanced photo filters, but now it involves creating entire images through a series of prompts, enabling pretty much anyone to generate any image they can imagine, independent of their skill set. However, this is not without limitations. Since current AI models do not “understand” what they generate, the results are frequently less than ideal outcomes such as characters with extra hands, fingers, or in unnatural poses. While performance has improved significantly compared to the first generative AI art engines, additional touch-up and verification is often required. Nonetheless, it’s an area that continues to evolve, and the models will continue to improve. For basic art game objects and items, this can be a time-saver in generating icons, tokens, or even characters for game development. It will be difficult to entirely replace artists, and I don’t believe it will, but it can assist artists in getting a head start and generate ideas quickly.

For example, if an artist wants to explore different art styles, generative AI can easily produce several samples, which can then be refined with some touch-ups based on feedback. In the realm of rapid prototyping, where low-fidelity visuals are common, AI-generated art will allow for faster iterations, enabling quicker decision-making to speed up game art production.

What about game design? Here, AI can also be helpful, as there are numerous game designs, theories, and models available to draw upon. Therefore, designing a difficulty level curve, a game coin economy with sources and sinks, and progression cadence can all be areas where AI can provide a great starting point for designers to begin ideating. An example would be a game designer feeding into an AI engine the rules of a combat system with various stats and then prompting it to simulate different battles to determine the outcome to help identify areas that require more balance. Will it be perfect? No, but it can be good enough to get started, as well as possibly surface other ideas that might have otherwise been missed. For very boot strapped studios it can help fill design holes for a proof of concept for a funding pitch or a game demo vertical slice to showcase a design concept.

Coding is another area rapidly benefiting from AI. Here, it’s like taking code libraries and putting them on steroids. Instead of taking an existing library and making tweaks as needed, you can now write a prompt to perform a task and have AI create customized code for that purpose all from scratch. This is great for basic tasks, but it can get quite sophisticated, where most of a basic game’s core could be coded up by a series of well-written prompts. For rapid prototyping development this can be a complete game changer. For QA and developers, AI can also be a great tool for scripting up various unit tests to check specific functions and mechanics in a game.

How about engagement? One area where AI has excelled is in language and text processing. You can find many examples of students using ChatGPT to compose entire essays. While games may not require essay writing, AI can be immensely helpful in crafting game flavor text and refining text for maximum engagement to captivate players. Need 500 engaging and concise push notification messages with a specific theme to send out to players? No problem! AI can help with that.

However, the real promise in engagement lies in using AI to predict user behavior and take corrective action automatically. For example, AI can predict that a particular player is likely to spend, and based on past behavior, it knows what types of boosters and items the player prefers, allowing it to craft highly personalized offers that the player is likely to purchase. Alternatively, if the AI detects that another player is struggling and likely to stop playing, it can offer personalized assistance to enhance the player experience and prevent churn. With AI machine learning it’s constantly learning and adjusting on what works and doesn’t work to engage the players and tune up or tune down the experience. These sophisticated examples require a multitude of AI techniques, but they highlight the enormous potential of AI in gaming engagement.

But it’s not all rosy; with all the potential of AI, change always brings disruption, both good and bad. For example, copyright and intellectual right questions have come into sharp focus. How can one determine if a generative art piece originated from a copyrighted image? Or if a generative art is made up of 100s of images and one happens to be copyrighted then potentially the entire art is usable and what would be considered “fair use”. Conversely, courts have already ruled (for now) that generative AI is copyrightable since “no human was involved”. This creates an interesting dilemma as anyone then can lift the same game art and use it for their purposed whether noble or nefarious.  Similarly, with code, how can one determine if some of the code was lifted from a proprietary source owned by someone else? This can certainly get very messy as AI navigates the legal system but it shouldn’t stop AI adoption in games. In particular, the potential around game engagement is enormous, where sophisticated AI models can learn and adjust based on player behaviors to create a highly tailored game experience that balances fun and delight for players while optimizing monetization. I believe AI adoption to games will continue to accelerate, and all game studios should be looking at how AI can be used to level up every aspect of game development – and I would love to hear your thoughts!

I believe AI adoption to games will continue to accelerate, and all game studios should be looking at how AI can be used to level up every aspect of game development.

Last updated on May 23, 2024