Veteran game writer David Gaider, best known for his work on the Dragon Age series, has expressed strong reservations about the current state of generative AI in game development, likening the technology to a "virulent plague." In a recent interview, Gaider argued that even when implemented to streamline mundane tasks, generative AI poses significant risks, particularly to the training and development of junior staff.
Gaider highlighted concerns regarding the data sources used to train generative AI models. He pointed out that the lack of control over these datasets "opens up any use of it to all sorts of future legal issues." This ambiguity surrounding data provenance could lead to unforeseen legal ramifications for studios employing the technology.
Beyond the legal uncertainties, Gaider questioned the practical utility of generative AI as a development tool. He suggested that relying on AI to edit or refine existing work, even to handle "drudgery," is often less efficient and yields poorer results than starting anew. "In all my time as a narrative designer I've never once encountered a situation where editing an inferior product took less time than simply throwing it out and redoing it would have or resulted in anything better than mediocre," Gaider stated.
A central argument from Gaider revolves around the potential elimination of entry-level tasks, which are crucial for onboarding new talent. "How are we going to train up the next generation of devs if we eliminate every entry-level task?" he questioned, emphasizing that these foundational roles provide essential learning opportunities for aspiring game developers.
Applying generative AI to programming presents similar challenges, according to Gaider. He expressed skepticism about creating prototypes with AI if the team does not gain practical knowledge of the development process. Similarly, he questioned the value of AI-generated concepts that may be "soulless and contain errors" and are not easily replicable by the studio's own artists. Furthermore, he raised concerns about adopting systems whose inner workings are not understood by the development team.
Gaider concluded that generative AI is "not ready for prime time," regardless of executive enthusiasm. He advocated for a cautious approach, recommending that the technology be avoided until it is properly regulated and its training data sources are clearly established and ethically sound. The developer views the current push for AI as an "AI gold rush" that overlooks the experimental nature and inherent risks of the technology.