Image by Victoria from Pixabay
We’ve come a long way from Pong. Since the arrival of video gaming in the middle of the 20th century, the industry’s advances have been constant and rapid.
Gameplay has become more sophisticated, graphics more lifelike, and the worlds we explore ever more immersive.
The last few decades in particular have seen the industry surge forward, largely driven by other technological breakthroughs.
There’s no shortage of coverage about what AI can do, and how it’s changing the world. Some of it is thrilling, like personalized healthcare, and some of it is terrifying, like AI impersonators, but the developments powering the 21st century’s AI revolution are empowering video game developers to produce incrementally more engaging gaming content.
- Generating characters
The players’ characters are no longer alone in the virtual world of video games. Non-playable characters (NPCs) are becoming increasingly complex and nuanced figures in the game’s narrative, and are a vital part of what makes a game engaging.
But generating characters is hard, and time-consuming. They’re easy to think of in the abstract but the details of each character are what bring them to life. These take time, but the advent of large language models offers an effective shortcut. Developers can use existing character biographies to prompt generative AI to produce more character outlines of the same kind.
Suddenly, it’s possible to generate thousands of characters, each with their own story and history, ready to populate the world around the player.
- Giving characters a voice
The development of natural language processing, which has played a huge part in the emergence of LLMs, also powers the ability to generate dialogue for NPCs.
This too can be developed from existing game dialogue, but again taking a fraction of the time. Instead of writing thousands of scripted lines, developers can rely on LLMs and generative AI to come up with responses to user actions.
Not only does this save time, but it also gets rid of the staid repetitiveness that was once characteristic of interactions with NPCs. The variation adds to the believability of the virtual world these characters populate.
- Making them meaningful characters
While this approach to character dialogue greatly improves the idle chatter players can experience, even allowing NPCs to talk to each other, generative AI also gives developers the ability to turn NPCs into characters in the narrative that can do things of consequence.
NPCs can say things that give the players tasks or actions, setting in motion the next part of the story, or creating a new narrative branch that allows for more novel gaming experiences.
This means LLMs and generative AI are not just making it easier to develop rich worlds full of believable characters, but they can actually contribute to the gaming experience of the player.
- Supporting developer workflow
As well as powering character development, LLMs are making the developers’ jobs easier too.
LLMs in combination with RAG systems can be used to power code assistants which can give feedback and guidance to developers in real-time as they’re writing code.
Far from threatening to put developers out of a job, this is freeing up time and headspace for developers to think about the next great innovation.
In combination, these aspects are revolutionizing video game development. Digital worlds will be fuller and bigger. The characters that live in them will have stories and lives that add to the space players are exploring. And they will be able to interact with players in a way that feels natural and believable.
But this is just the start of this next step in video game development. Gaming tends to jump forward in fits and bursts as new technologies arrive and the industry internalizes them. As more and more developers learn how to get the most out of LLMs and generative AI in their development practices, the results are only going to get more impressive.