This article, part of AI Week, features a discussion between Daniel Griliopoulos, a journalist turned games writer known for work on Total War: Warhammer 3 and co-author of Ten Things Video Games Can Teach Us About Life, and Thomas Keane, co-founder of Meaning Machine, creators of the AI-powered game Dead Meat. They debate AI’s role in game narratives, addressing ethical issues, creative potential, and industry responsibilities. On current views of AI, Daniel highlights significant ethical concerns including power imbalance, plagiarism, and job losses, noting that while ethical AI is possible, current tools often ignore morality. He observes widespread use of AI outside the gaming sector. Thomas agrees with the concerns but stresses the importance of creatively and ethically embracing AI within the gaming community to lead its development responsibly rather than avoiding it. He worries about a lack of accountability in AI’s adoption. Regarding AI’s role in narrative, Thomas argues that AI alone generates poor-quality, incoherent content; meaningful storytelling requires human input. Meaning Machine uses AI not to replace but to empower human creativity, employing it at runtime to offer unprecedented flexibility, adaptability, player freedom, and self-expression. Daniel agrees AI enables new possibilities but believes design, largely human-driven, mainly creates valid experiences. He points out that human creativity inherently involves borrowing and adaptation, much like AI's training on existing materials, raising ongoing ethical concerns around plagiarism and power. He notes current AI-driven narratives lack the reliability and quality of hand-curated content, which remains the preferred approach for better player experiences. Thomas agrees AI storytelling quality is currently insufficient but aims to improve it by blending AI with handcrafted, copyright-free content. Their system “bullies” AI into producing more interesting, specific outputs by grounding AI responses in detailed human-authored narratives, creating unique and engaging stories. Both emphasize the value of constraints; true open-ended AI conversations often generate irrelevant or frame-breaking dialogue, so guiding AI through tight parameters is necessary for quality. On AI NPCs able to talk about anything, Daniel notes existing AI still struggles with reliable constraints and quality, meaning handcrafted stories by experienced writers remain superior. Thomas insists their method can match handcrafted content quality while adding flexibility and points to independent research by the University of Bristol confirming players appreciate the creative freedom and immersion their AI-driven narratives provide. He stresses that AI is not meant to replace traditional storytelling but to broaden gaming’s narrative possibilities, much like different gaming formats coexist without replacing each other. Looking ahead, Daniel suggests AI may expand procedural content like side quests or incidental dialogue but is unlikely to penetrate main storylines soon due to ethical concerns.

Thomas sees AI as a driver of gaming’s ongoing radical reinvention—augmenting rather than replacing existing forms. He notes many reputable developers are quietly experimenting with AI storytelling behind the scenes, anticipating a future wave of improved quality and acceptance. On concerns about AI being cheaper than human labor, Daniel warns of hidden external costs such as environmental impact and ethical trade-offs. Thomas echoes this, emphasizing power-efficient, on-device AI use and stresses that their approach requires more human writers to craft the foundational content that the AI then elaborates on. He calls this the “era of the writer, ” where human-authored material remains central to quality, meaning AI is not a cheap shortcut but a different kind of creative collaboration. Daniel distinguishes between narrative design and writing in games, suggesting AI may reduce traditional writing jobs while elevating narrative design roles. He personally enjoys the act of writing itself and prefers to create directly rather than curate AI outputs, comparing it to preferring to paint a picture over generating it through AI art tools. Thomas argues their approach costs about the same overall but shifts effort towards authoring extensive handcrafted content to fuel emergent AI storytelling. He compares the transition to how 3D graphics revolutionized games, requiring new skills but enabling fresh experiences. On ethics, Daniel sees potential for ethical AI if copyright reform occurs, such as retraining models on non-copyrighted data or compensating creators for model use (e. g. , voice actors licensing their voices). Power consumption issues might be mitigated through efficient on-device AI. Job losses are a wider societal problem needing government support for those affected. Thomas concurs and emphasizes legislation must evolve to ensure fair compensation and sustainable AI development, advocating for active human roles in AI-assisted creation to prevent workforce eradication by corporate interests. Both agree there is an ethical imperative for the games industry to engage proactively with AI to shape its impact positively, lest the worst actors dominate. In conclusion, the discussion highlights AI’s promise as a tool for innovative, player-driven narratives when ethically integrated with human creativity and oversight. While challenges remain in quality, ethics, and job impacts, thoughtful adoption can enrich gaming’s storytelling landscape without sacrificing artistic integrity or fairness. AI is positioned not as a replacement but as a new creative partner that will coexist with traditional narrative practices, driving forward the evolution of game design. (Note: This interview has been edited for length and clarity. )