獨立遊戲開發者對人工智慧技術接受行為之研究

Abstract

近年來，生成式人工智慧技術迅速發展，對全球文化創意產業產生了深遠的影響，其中也包含了遊戲開發產業。然而，過去相關研究多半聚焦於大型遊戲公司，較少探討獨立遊戲開發者的情況。本研究旨在填補此研究缺口，深入探討生成式AI對於台灣獨立遊戲開發者的影響，尤其是在開發流程、創意表現、市場競爭力以及內部組織人力方面的改變。本研究使用質性研究中的多個案研究法，針對三間不同類型的台灣獨立遊戲開發工作室與兩間產業相關協會進行深度訪談。研究以整合科技接受模型（TAM）與技術－組織－環境（TOE）理論的TAM-TOE模型為分析框架，系統性地分析開發者對於生成式AI技術的採用行為與其背後的脈絡因素。研究發現，台灣獨立遊戲開發者普遍將生成式AI視為提升效率的輔助工具，應用於概念發想、資料蒐集、美術風格測試、程式編碼輔助、文本翻譯及次要內容生成等多個環節。AI的導入有效加速了遊戲迭代速度，在一定程度上彌補了小型團隊的技能短板並降低了部分重複性工作的成本。然而，開發者也面臨了諸多挑戰，包括AI生成內容的品質不穩定、著作權歸屬的法律爭議，以及玩家社群對「AI感」的強烈反感。為此，開發者的應對策略傾向於將AI用於內部或不易察覺的環節，並投入人工進行後製修改與品質把控。在組織人力方面，AI的出現取代了部分外包與初階工作，但也催生了人機協作的新模式與對AI工具審核、優化等新技能的需求 。本研究結論指出，台灣獨立遊戲開發者正處於導入AI的探索與適應期。開發者在享受AI帶來效率優勢的同時，也需要謹慎應對其技術限制、法律風險與市場反應，在人機協作中尋求平衡點，以維持團隊的創作活力與市場競爭力。

This study investigates the impact of Generative AI on Taiwanese indie game developers, an area where research has been sparse. Adopting a qualitative multiple case study approach, the research conducts in-depth interviews with three indie game studios and two industry-related organizations. The Technology Acceptance Model-Technology-Organization-Environment (TAM-TOE) framework is utilized to analyze how developers perceive and integrate AI, considering technological, organizational, and environmental factors.The researches reveal that developers primarily use Generative AI as an auxiliary tool to enhance efficiency rather than as a replacement for human creativity. Its application is most common in the early and supporting stages of development, including ideation, art style prototyping, programming assistance, and generating non-core audio content. This integration has begun to transform traditional workflows by accelerating game iteration and reducing pre-production communication costs, particularly benefiting small teams by helping to bridge skill gaps.Despite the benefits, developers face significant challenges, including the inconsistent quality of AI outputs, unresolved copyright issues, and strong player resistance to an AI feel in games. To mitigate these issues, developers employ cautious strategies, using AI for internal processes or less visible content and relying heavily on manual review and refinement to ensure quality. Consequently, while AI has replaced some routine outsourced tasks, it has also spurred a shift in job roles toward human-AI collaboration, creating new demands for skills in managing and optimizing AI-generated content.