以開源軟體進行知識轉移之探討: 以 D8AI 公司為例

Abstract

本研究主要目的是深入研究人工智慧服務公司D8AI以研發開源軟體進行知識轉移而達成公司績效目標。將探討企業為何選擇導入人工智慧服務的動機和背後的推動因素及如何創造更多價值，企業內部運營和供應鏈合作關係中的實際應用，協助企業更有效地應對人工智慧服務導入的相關經驗，最終為企業決策者和業界利益關係者提供人工智慧領域的實用建議。人工智慧最顯著的趨勢是技術的普及和開放，人工智慧技術不再僅限於科學實驗室，而是進入了各個產業，促使了更多創新的產品和服務整合專業技術加速了創新，因此建構出以開源軟體進行知識轉移架構是能影響創新績效的因素。本研究訪談了D8AI公司以及進行知識轉移的AI推廣方案合作夥伴，透過質性研究，深入瞭解與探討開源軟體進行知識轉移的過程。本研究得到的主要討論有：（1）開源軟體的快速迭代與更新使企業能夠及時獲取最新技術，且能降低企業的研發人力成本和軟體費用，加速了流程與產品的創新，促進企業績效；（2）D8AI選擇研發合作夥伴的因素包含資源及技術的互補（例如：大型語言模型與語音辨識）與財務能力，合作的動機會因核心能力、研發策略的差異而有所不同；（3）D8AI選擇與第三方機構或大學研究機構進行研發合作的過程中會簽訂保密協定，並以不具直接競爭關係者為主要合作對象，也以政府機構的合作機會為目標，提升彼此競爭力且不影響商業模式；（4）D8AI以自然語言處理為研發核心並培養AI人才，且透過國際性電腦展覽的參與來行銷技術與產品、接觸客戶與吸收市場資訊，未來發展應用將囊括各行各業，以台灣人工智慧產業翹楚為目標。並據於此，本論文最後提出三項實務建議及後續研究建議，以供未來相關領域的廠商及研究者作為參考。

The main purpose of this study is to delve into how the artificial intelligence service company D8AI achieves its performance goals through the development of open-source software for knowledge transfer. It will explore the motivations and driving factors behind enterprises' choice to adopt AI services, how they create more value, and the practical applications within internal operations and supply chain partnerships. This assists enterprises in more effectively coping with the experiences related to the introduction of AI services, ultimately providing practical advice in the AI field for enterprise decision-makers and industry stakeholders. The most notable trend in artificial intelligence is the popularization and openness of the technology. AI technology is nolonger confined to scientific laboratories but has entered various industries, promoting the integration of more innovative products and services with professional technology, thus accelerating innovation. Therefore, constructing a framework for knowledge transfer through open-source software is a factor that can influence innovation performance.This study interviewed the company D8AI and its partners in the AI promotion program involved in knowledge transfer. Through qualitative research, it deeply understands and explores the process of knowledge transfer through open-source software. The main conclusions of this study are:(1)The rapid iteration and update of open-source software enable enterprises to acquire the latest technology in a timely manner, reduce the cost of R&D manpower and software expenses, accelerate the innovation of processes and products, and promote enterprise performance;(2)The factors that D8AI considers in selecting R&D partners include resource and technical complementarity (e.g., large language models and speech recognition) and financial capability. The motivation for cooperation varies due to differences in core competencies and R&D strategies;(3)During the process of R&D cooperation with third-party institutions or university researchinstitutions, D8AI will sign confidentiality agreements and mainly cooperate with those who do not have direct competitive relationships, targeting opportunities for cooperation with government agencies to enhance mutual competitiveness without affecting business models;(4)D8AI focuses on natural language processing for R&D, nurtures AI talents, participates in international computer exhibitions to market technology and products, engage with customers, and absorb market information. Its future development applications will cover various industries, aiming to be the leader in Taiwan's AI industry.Based on this, the paper finally proposes three practical recommendations and suggestions for follow-up research for reference by future related industry practitioners and researchers.