生成式人工智慧應用於食農知識創新服務與管理之研究

Abstract

本研究「生成式人工智慧應用於食農知識創新服務與管理」旨在探討生成式人工智慧如何應用於食農產業的知識管理，及在臺灣既有食農知識體系之窠臼下，如何善用此破壞式創新科技加速數位轉型，以及專業知識服務如何以此槓桿借力使力提升領域知識與服務能量以支持食農產業創新。本研究深度訪談產、學、研之學者專家、進行文本分析與開放性、軸向性、選擇性之三階段編碼、並整合知識圖譜視覺化驗證技術，以多角度、多層次探討生成式AI（尤其是大型語言模型）應用於臺灣食農知識體系之現況與挑戰，並以某食農技術服務業者為案例，示範企業知識圖譜建模，及其整合檢索增強技術（RAG）應用於知識管理以降低AI幻覺，並提升專業度之實務價值，並進行多元資料交叉分析及比對。歸納出本研究核心現象之理論模型 - 「AGILE by AI：敏捷智能組織互動模型」以貫穿整個研究，並說明生成式AI應用於食農知識創新服務與管理之施行要項如下： （一） 生成式AI作為提升食農知識服務與管理之新動能。 （二） 加速臺灣食農知識體系之數位化與創新進程。 （三） 知識圖譜是提升生成式AI在食農產業知識管理應用可靠度之關鍵。 （四） 臺灣食農產業導入生成式AI需考量其產業特性與挑戰。 （五） 人機協作是發揮生成式AI在知識管理之最有價值與有效途徑。

基於以上核心主題，本研究以生成式AI透過與知識圖譜的有效整合，並在人機協作的架構下，提升食農產業的知識創新與服務管理效能，解決產業長期面臨知識分散、傳承困難等問題，進而促進臺灣食農知識體系的數位轉型與智慧升級，但在導入過程中需充分考量產業的特性與挑戰，積極應對潛在的風險。此分析結果呈現於文獻探討中關於生成式AI應用於食農創新服務與知識管理研究的主要面向和核心觀點。後續的研究基於此進行更深入的探討與驗證，以歸納具理論基礎之實證架構，為未來相關領域的應用與研究提供參考依據，並期望成為食農中小企業未來應用或導入生成式AI於知識管理之系統性指引。

This study aims to explore how Generative Artificial Intelligence (GenAI) can be adopted on Knowledge Management (KM) for Agrifood Knowledge Innovation Services, and to analyze benefits and risks of such deployment. The research method utilizes multiple approaches of Qualitative Analysis, starts with Literature Review, then by Textual Analysis following In-Depth Interviews of stakeholders of Agrifood industries and experts in GenAI. As the third pillar, taking an Agrifood Tech company for case-study, analyzing its corporate and industrial knowledge involved with Agrifood industrial value chain. Furthermore, cross-indexing extracted concepts and thematic entities by constructing Knowledge Graph (KG), which can be further utilized as a Visualization Tool of respected contexts of known or unknown knowledge, and a Validation Technique for finetuning GenAI outputs for knowledge creation and management. Ultimately, to draw a conclusion with the “AGILE” Model, which is the acronym to best describe the Model as “AI driven knowledge Grounding Interactions Leveraged by workforce Empowerment”. The key issues derived from the study are as follows: 1. GenAI as a Technical Driver for enhancing knowledge creation, sharing and management for industries. 2. GenAI is the Engine for Digital Transformation of Taiwan existing Agrifood Knowledge Innovation System (AKIS). 3. Knowledge Graphs as the Key to improve the reliability of GenAI implementation in Knowledge Management within organization. 4. Industry-specific Challenges to be considered for adoption of GenAI in Taiwan's Agrifood sector. 5. Human-AI Collaboration as an effective approach to maximizing the benefits of GenAI in knowledge management maneuver. Based on these key issues, the study anticipates, by integrating Knowledge Graphs and leveraging Human-In-The-Loop Configuration, the positive circulation of knowledge creation can be greatly promoted. This collaboration aims to address long-standing challenges such as dispersive and fragmented natures of Business Intelligence and difficulties in communicating and sharing. The anticipated findings will contribute to the literatures outlining key perspectives and insights on GenAI and Knowledge Management within the Agrifood sector, Future studies can build upon these grounds for further exploration. Additionally, it aims to offer a systematic guideline for small and medium-sized enterprises seeking to adopt GenAI for boosting productivity and creativity. However, it’s highly dependent to the leaderships who wish to integrate GenAI into their workflows with a mindset of transformation both digitally and organizationally.