1 free read left. Create a free account for unlimited reading + AI research.
Create free accountWhitepaper
DiGRA
The paper argues that generative AI will play a major role in videogame development and gameplay. Existing regulatory frameworks do not fully address the ethical challenges its use introduces.
This paper, authored by Jana Hecktor and Theresa Krampe at the University of Tübingen, aims to develop ethical guidelines for using generative AI (GenAI), including large language and multimodal models, in the creation of videogame characters. Its central argument is that GenAI is poised to play a major role in game development and gameplay, yet its use introduces ethical challenges that existing regulatory frameworks do not fully address. The authors focus on two applications they consider most promising: character model creation and conversational AI for non-player characters (NPCs). The work is positioned as a value-sensitive approach to game design, intended to help designers make informed decisions about when and how to deploy GenAI.
The analysis proceeds from a review of existing AI ethics frameworks and identifies four core values to guide the approach: fairness, sustainability, transparency, and privacy protection. Fairness is framed both as the just distribution of AI's costs and benefits and as the mitigation of discriminatory bias, extending to labor conditions such as "click-work" in the Global South and to the representation of player communities. Sustainability is treated primarily in its environmental sense, given the energy and resource demands of training and running GenAI. Transparency concerns the explainability of systems to designers and players, including disclosure when players are interacting with an AI, while privacy addresses the substantial volume of sensitive player data that gaming devices already collect and that GenAI may intensify.
The paper also reports a preliminary, exploratory interview study with professionals who use or plan to use GenAI for NPC design, analyzed using grounded theory methods. Participant numbers are not specified in the text. The interviews surfaced perspectives not prominent in the literature, such as designers' sense of solidarity with fellow artists regarding copyrighted training data and the observation that ethical requirements vary by game genre. A second, refined interview round is planned, with the final stage consolidating theoretical and empirical findings into a blueprint for industry guidelines.
PocketGamer.biz · 2023