Jam & Tea Studios, a new video game startup, is using generative AI technology to redefine how players interact with non-playable characters (NPCs) in video games.
This innovative approach aims to transform player engagement by going beyond traditional scripted NPC behavior, which often feels monotonous and unrealistic. By integrating AI, Jam & Tea enables more dynamic and personalized interactions, allowing players to have natural conversations with NPCs and explore a wider range of experiences.
Founded by veterans of Riot Games, Wizards of the Coast, and Magic: The Gathering, Jam & Tea recently announced its first game, Retail Mage, which will use generative AI to enhance various aspects of gameplay. The game will use AI tools to manage game mechanics, generate content, create dialogue, and even produce items, expanding the possibilities of the game world.
Retail Mage is a role-playing game (RPG) in which players take on the role of a wizard working as a salesperson in a magical furniture store. The primary goal is to earn five-star customer reviews, though players can choose to either diligently help customers or cause chaos. With AI-based NPCs acting as customers, the game opens up a wide range of potential outcomes based on player choices and interactions.
In Retail Mage, customers make unique requests to players, and instead of relying on predefined dialogue options, players can type their responses into a text generator. This allows players to type in commands like “say something charming,” prompting the AI to generate multiple dialogue options in real-time. The inclusion of AI-driven NPCs promises to make the gameplay experience more spontaneous and engaging.
Jam & Tea isn’t alone in exploring AI-enhanced NPC interactions. Other companies like Artificial Agency, Inworld, and Nvidia are also developing similar technologies. Additionally, major gaming companies like Ubisoft have introduced AI-powered tools like “Ghostwriter” to generate dialogue for NPCs in their games.
While generative AI offers significant benefits, it also presents challenges. One concern is AI unpredictability, where NPC behavior can become erratic, leading to a frustrating player experience. There is also a risk of AI “hallucinations,” in which NPCs may provide inaccurate or nonsensical responses. To address these issues, Jam & Tea plans to continually improve its AI engine and implement safeguards to prevent inappropriate conversations. Players can also rate NPC responses, providing valuable feedback to refine character behavior.
The game encourages creativity, allowing for inventive interactions that can lead to unexpected scenarios. For example, during a playtest, a player expressed boredom, prompting an NPC to suggest a game of hide-and-seek. This spontaneous activity was not pre-programmed but emerged naturally from the AI’s ability to adapt to player input, demonstrating the potential of generative AI to create engaging experiences in real time.
Jam & Tea has experimented with various large language models (LLMs), including OpenAI, Google’s Gemma, Mistral AI, and Meta’s Llama, and is currently refining its chosen model to improve character responses.
Beyond dialogue, the AI engine in Retail Mage extends to object interaction, allowing players to manipulate or create items based on their intentions. In a demo, players could collect or craft items, such as summoning a stuffed antelope pillow for an NPC customer. While the physical item may not appear visually, the action is recognized in the game’s inventory system, providing a canvas for the player’s imagination and creativity.
Jam & Tea ensures that AI technology will not replace the work of artists, as all 2D and 3D elements in the game will be created by real human designers. This commitment highlights the studio’s balanced approach to integrating AI while preserving the integrity of creative contributions.
With just eight team members, Jam & Tea faces the challenge of competing with the largest gaming companies. However, by pioneering AI technologies early on, the studio is positioning itself to adapt and grow alongside advances in AI models. The company has already secured $3.15 million in seed funding and plans to raise additional capital to support its growth.
Retail Mage will be available for $15, with additional game packs available for purchase. Initially launching on PC, the company plans to expand to cross-platform compatibility in the coming years. The game is expected to be released to the public later this fall.
Écrit par Vytautas Valinskas
Originally published in The European Times.
source link eu news
