| 摘要: | 近年來快速發展的人工智慧（Artificial Intelligence, AI），其特徵是以大數據、雲計算以及深度學習等三個工具為基礎，並成功應用在語音辨識、圖形識別以及自然語言處理等領域，進而讓自動駕駛、語音助理等技術獲得突破性的進展，這一波人工智慧的發展勢必將讓人類生活產生深遠影響。淘寶網藉由深度學習技術整合阿里集團的大數據資料與阿里雲的雲端運算服務，於2017年推出AI客服聊天機器人「店小蜜」，提供消費者一種全新的溝通方式。「店小蜜」一方面提升了電子商務的運作效率，同時也讓人與人工智慧共同合作的生態提早到來。因此，探討企業或消費者對此類以人工智慧為基礎之創新服務的需求與選擇行為，對於學術界或產業界而言，都是一個極重要且具挑戰性的課題。本研究為兩年期計畫，第一年是以台北、上海與香港等地的大學生消費族群為研究對象，藉由整合個體選擇模型與劇變模型來探討影響消費者在網路購物的不同階段中，對於AI客服的需求與其選擇行為。考慮到選擇行為本身的不連續變化現象，本計畫將藉由建構蝴蝶劇變模型來探討選擇AI客服行為的非線性特徵，並挑選會讓選擇產生不連續變化的變數來修正個體選擇模型，進一步分析在不同購物階段情境下，消費者對AI客服需求的不同認知與選擇型態。第二年的研究計畫是以淘寶網與中國亞馬遜書店為研究個案，經由模糊認知圖與敏感度模式等分析技術，導入脆弱度、回復力與反脆弱的觀點建構研究個案開發以AI為基礎的商務服務時，相關風險管理的分析架構與評估機制。計畫的研究成果期能深入理解影響消費者與企業選擇創新服務的因素外，並藉由本研究的理論模式所衍申出之管理意涵，提供相關業者營運規劃之參考依據。
In recent year, the rapid development of artificial intelligence is based on the big data, cloud computing and deep learning tools such as the foundation, and successfully used in speech recognition, pattern recognition and Natural Language Processing and other fields, and then get breakthrough progress for automatic driving, voice assistant technology, the development of this wave of artificial intelligence will make human life have a profound impact. Taobao has integrated big data, cloud computing services and deep learning technology, and launch of the AI customer service chat robot "Dian Xiaomi" in 2017. On the one hand, the service improves the efficiency of the operation of e-commerce, and it also brings the ecological cooperation between people and artificial intelligence early. Therefore, it is an important and challenging topic to explore the demand and choice behavior of such innovative services based on AI for enterprises or consumers. To better understand the crucial factors of behavior intention in use of AI service, this study is intended to investigate what kinds of factors effect of using AI service on consumers' attitude towards in e-commerce market. The research approach combined with logit model and butterfly catastrophe model. From the survey, we can explore what kind of factors can influence AI service choice behavior intention via discrete choice model and butterfly catastrophe model. Risk management is regarded as the important issue in e-commerce, it is no doubt if managers could capture the consumers' feature and focus on the management risk issue, it will increase the competitiveness of enterprises. In this study (second year), we are concerned about Taobao and amazon (China) and their AI service. The objective of the second year in this study is to develop an evaluation model and discuss the risk of AI service system via the Fuzzy Cognitive Map (FCM) and Sensitivity Model (SM). From the survey, we can establish an evaluation model to analyze and describe the antifragile, vulnerability and resilience of the AI service system using different kind of research methods. The results obtained in this study can be used to improve the AI service quality for e-commerce and development more effective management strategies. |
