Language

AI tool for auto-generation of the social and narrative dimension of a game universe

Project partners

 

The game studio GenNarrate, in collaboration with Aalborg University (AAU), is developing an artificial intelligence that makes it possible to autogenerate the social and narrative dimension of a game universe. The technology must simulate human behavior of Non-Player Characters and enable characters in computer games to make free and emotion-driven choices to achieve a specific goal. The technology is similar to the existing tools that autogenerate a computer game’s physical environment.

 

The collaboration with Aalborg University is a partial phase in the development of the AI ​​tool Genarator, and must search different AI algorithms and find the most appropriate ones depending on which characteristics the game characters must have (silly, smart, fast, likable).

Facts

Started: January 2021

Completed: June 2022

Problem
The business model in game development has moved from one-time payment solutions to subscription schemes (software as a service), where players can play as much as they want for a fixed monthly amount. This requires the game developer to constantly develop their game to keep the players engaged, so that they get new experiences every time they play. A high retention makes the game more attractive to game publishers and potential investors.

But it takes time and costs money to continually develop new content for your game.

 

Solution
With help from Aalborg University, GenNarrate is working on developing a new type of content generation tool that autogenerates the social and narrative aspect of a computer game. It will make it easier and faster to develop a computer game’s narrative, provide a more authentic game experience and thereby increase the game’s retention (time spent on the product).

 

Result
The collaboration with AAU contributes a sub-element, an algorithm, of the Generator technology. It will result in a demo game in which you can test the technology. The results of the project, including the algorithms, were presented at the international conference ICAPS 2022 workshop and are now publicly available here.

Interested in hearing more about the project? contact:

Susanne Skovgaard, Vision Denmark
E-Mail: ssc@visiondenmark.dk

Powered by