社群媒體上的生成式人工智慧輿情探勘;Public Opinion Mining of Generative AI

NCUIR > School of Management at National Central University > Graduate Institute of Information Management > Electronic Thesis & Dissertation > Item 987654321/95482

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/95482

Title:	社群媒體上的生成式人工智慧輿情探勘;Public Opinion Mining of Generative AI
Authors:	陳敬元;Chen, Ching-Yuan
Contributors:	資訊管理學系
Keywords:	生成式人工智慧;社群媒體;輿情分析;情感分析;隱含狄利克雷分布;情感分類;Generative AI;Social media;Opinion mining;Sentiment analysis;LDA;Emotion Classification
Date:	2024-07-10
Issue Date:	2024-10-09 16:53:46 (UTC+8)
Publisher:	國立中央大學
Abstract:	生成式人工智慧 (GAI) 是當前發展最快且廣受討論的技術之一，對社會產生了重大影響。生成式AI能夠產生大量豐富且有趣的內容，但也因為一些問題而受到批評，例如侵犯版權或智慧財產權。本研究旨在利用自然語言處理技術探討社群媒體用戶對生成式AI內容的看法和情緒反應。研究聚焦於社群媒體平台Reddit，深入收集和分析用戶的討論和意見。數據範圍涵蓋2023年7月至2024年4月，來自多個子板塊和關鍵字，涵蓋文字、圖片和影片生成等各種生成式AI技術。本研究使用RoBERTa預訓練模型進行兩階段情感分類，分析貼文和評論中表達的情緒，並將其分類為主動和被動情緒。研究還使用潛在狄利克雷分佈來識別主要的討論主題，探討用戶對不同類型生成式AI技術的主要關注點。研究結果表明，只有相對較低比例的用戶對生成式AI技術表達了負面情緒，這表明即使GAI引發了許多爭議，大多數公眾仍對其保持積極的態度。雖然使用者對各種類型生成式AI技術的主要關注點略有不同，也存在許多相似的討論話題。本研究為技術開發者、政策制定者和社群媒體平台管理者提供了了解公眾情緒和解決生成式AI相關問題的看法。;Generative artificial intelligence (GAI) is one of the fastest developing and widely discussed technologies, having a significant impact on society. However, while generative AI can produce a lot of rich and interesting content, it has also been criticized for some issues such as violating copyright or intellectual property rights. This study aims to explore social media users′ views and emotional responses to generative AI content using natural language processing techniques. The research focuses on Reddit, a social media platform, by deeply collecting and analyzing user discussions and opinions. The data spans multiple subreddits and keywords from July 2023 to April 2024, covering various generative AI technologies such as text, image, and video generation. This study employs the RoBERTa pre-trained model for a two-stage sentiment classification to analyze the emotions expressed in posts and comments, categorizing them into active and passive sentiments. Additionally, the study uses Latent Dirichlet Allocation to identify the main discussion topics, exploring users′ primary concerns regarding different types of generative AI technologies. The study′s findings indicate that a relatively low proportion of users expressed negative sentiments toward generative AI technology. This suggests that despite the numerous controversies surrounding GAI, the majority of the public maintains a positive outlook on it, and although there are slight differences in the main concerns regarding various types of generative AI technologies, there are also many similar discussion topics. This research provides insights for technology developers, policymakers, and social media platform managers to understand public sentiments and address concerns related to generative AI.
Appears in Collections:	[Graduate Institute of Information Management] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	16	View/Open

社群 sharing

Loading...