頁籤選單縮合
| 題 名 | AI與學術期刊:賦能與祛魅=AI and Academic Journals: Empowerment and Disenchantment |
|---|---|
| 作 者 | 張耀銘; | 書刊名 | 澳門理工學報. 人文社會科學版 |
| 卷 期 | 28:4=100 2025.10[民114.10] |
| 頁 次 | 頁125-141+242 |
| 分類號 | 312.83 |
| 關鍵詞 | 生成式人工智能; 三大學派; 人機協作; AI幻覺; 版權侵權; 倫理風險; Generative artificial intelligence; Three major schools; Human-machine collaboration; AI hallucinations; Copyright infringement; Ethical risks; |
| 語 文 | 中文(Chinese) |
| 中文摘要 | 1956 年達特茅斯會議標誌著人工智能的誕生,符號主義、連接主義與行為主義三大學 派共同構建了 AI 的理論基石。 歷經 70 年“兩起兩落”的技術浪潮,生成式人工智能( GAI) 在 21 世 紀實現了從“分析系統”到“生成系統”的範式躍遷,推動人工智能從“模仿人類”走向“ 協同人類” 的 “後學派時代”。 生成式人工智能對學術生態的影響呈現多維變革:在科研範式維度,催生“ 第五科 研範式”,AlphaFold 蛋白質預測、AI 輔助定理證明等案例,打破了傳統學科壁壘,實現數據驅動與模 型驅動的融合;在知識生產維度,從主題構建、文獻綜述到文本修飾,AI 以高效的信息整合能力輔助 作者重構創作模式;在編輯審稿維度,學術期刊編輯必須提高識別 AI 生成論文的能力,從編輯加工 角色進化為“人機協作管理者”,通過一次次與 AI 的合作,實現熱點追蹤、選題策劃、稿件審讀、同行 評審等工作效率的提升。 然而,技術狂飆背後暗藏深層風險:訓練數據的版權模糊性、AI 生成物的 著作權歸屬爭議、“AI 幻覺”導致的虚假信息傳播,均對現有學術規範與法律體系提出挑戰。 如何 在技術賦能中守護學術本真,在範式變革中堅守人文內核,不僅是學術期刊面臨的時代命題,更是 整個知識生產領域需要共同書寫的答卷。 |
| 英文摘要 | The 1956 Dartmouth Conference marked the birth of artificial intelligence (AI), with three major schools—symbol- ism, connectionism, and behaviorism—contributing to its theoretical foundations. Over the past 70 years, characterized by cycles of technological rise and fall, generative artificial intelligence (GAI) has achieved a paradigm shift from “ analytical systems” to “ generative systems” in the 21st century, moving AI from “ mimicking humanity” to “ collaborating with hu- manity” in what can be termed the “ post-academic era.” The impact of generative artificial intelligence on the academic e- cosystem manifests as a multidimensional transformation: in the dimension of research paradigms, it has given rise to a “ fifth research paradigm”, exemplified by cases such as AlphaFold protein prediction and AI-assisted theorem proving, which break down traditional disciplinary barriers and integrate data-driven and model-driven approaches. In the dimension of knowledge production, AI assists authors in reconstructing creative processes through efficient information integration, from topic construction and literature reviews, to text refinement. In the editing and review dimension, academic journal edi- tors must enhance their ability to identify AI-generated papers, evolving from traditional editorial roles to “ human-machine collaboration managers” who improve efficiency in hotspot tracking, topic planning, manuscript review, and peer evaluation through repeated collaboration with AI. However, beneath this technological surge lie profound risks: the ambiguity of copy- right in training data, disputes over the authorship of AI-generated content, and the spread of misinformation due to “ AI hal- lucinations” all challenge existing academic norms and legal frameworks. The question of how to safeguard academic au- thenticity amidst technological empowerment, while maintaining a humanistic core during paradigm shifts is not merely a timely issue for academic journals, but also a collective challenge that the entire knowledge production field must address. |
本系統中英文摘要資訊取自各篇刊載內容。