人工智能生成内容中的话语认知分析
Cognitive Analysis of Discourse in AI-Generated Content
摘要: 本研究探讨了人工智能(AI)生成内容中的话语认知特征，旨在分析这些内容在语言表达、认知负荷、语义连贯性等方面的特性与差异。通过系统文献回顾，概述了AI生成模型的发展历程和技术基础，特别是对GPT-3等先进模型进行了深入探讨。研究采用质性与量化分析相结合的方法，从语言复杂度、情感色彩和认知负荷等多维度，对比分析了AI生成文本与人类生成文本。实验结果显示，AI生成内容在语言一致性和逻辑连贯性方面表现优秀，但在语义深度和情感表达上存在一定局限。进一步分析表明，这些差异可能源于算法设计、训练数据和生成机制的不同。本文提出了提升AI生成内容质量的建议，并为未来人工智能语言模型的优化和应用提供了科学依据和参考。
Abstract: This study explores the cognitive features of discourse in content generated by Artificial Intelligence (AI), aiming to analyze the characteristics and differences in terms of language expression, cognitive load, and semantic coherence. Through a systematic literature review, the development and technical foundations of AI-generated models are outlined, with an in-depth discussion of advanced models such as GPT-3. Employing a combination of qualitative and quantitative analysis methods, the study compares AI-generated texts with human-generated texts across multiple dimensions, including language complexity, emotional tone, and cognitive load. Experimental results indicate that AI-generated content performs well in linguistic consistency and logical coherence, yet shows certain limitations in semantic depth and emotional expression. Further analysis suggests these differences may stem from variations in algorithm design, training data, and generation mechanisms. The paper proposes recommendations for improving the quality of AI-generated content and provides scientific evidence and references for the optimization and application of future AI language models.
文章引用：沈越, 赵振强. 人工智能生成内容中的话语认知分析[J]. 交叉科学快报, 2024, 8(3): 238-247. https://doi.org/10.12677/isl.2024.83029

参考文献

[1] Brumby, D. (2004) Artificial Intelligence: A Modern Approach. Telecommunication Journal of Australia, 54, 63-64.
[2] Nilsson, N.J. (1996) Artificial Intelligence: A Modern Approach. Artificial Intelligence, 82, 369-380. [Google Scholar] [CrossRef] 
[3] Cao, C., Liu, F., Tan, H., Song, D., Shu, W., Li, W., et al. (2018) Deep Learning and Its Applications in Biomedicine. Genomics, Proteomics & Bioinformatics, 16, 17-32. [Google Scholar] [CrossRef] [PubMed]
[4] Seo, J., Moon, H., Lee, C., Eo, S., Park, C., Kim, J., et al. (2022) Plain Template Insertion: Korean-Prompt-Based Engineering for Few-Shot Learners. IEEE Access, 10, 107587-107597. [Google Scholar] [CrossRef] 
[5] Sun, G., Liang, W., Dong, J., Li, J., Ding, Z. and Cong, Y. (2024) Create Your World: Lifelong Text-to-Image Diffusion. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46, 6454-6470. [Google Scholar] [CrossRef] 
[6] Child, G.S. (2019) Book Review: James Paul Gee, an Introduction to Discourse Analysis: Theory and Method (Fourth Edition). Qualitative Research, 19, 236-238. [Google Scholar] [CrossRef] 
[7] Halliday, M.A.K. and Hasan, R. (1976) Cohesion in English. Longman.
[8] Halliday, M.A.K. (2004) An Introduction to Functional Grammar. Hodder Education Bookpoint Limited.
[9] Fairclough, N. and Nekvapil, J. (2006) RECENZE—Analysing Discourse. Textual Analysis for Social Research. Socio-logicky Casopis, 65, 435-438.
[10] Lakoff, G. and Johnson, M. (2003) Metaphors We Live By. University of Chicago Press. [Google Scholar] [CrossRef] 
[11] The Linguistic Society of Korea (1982) Linguistics in the Morning Calm. Hanshin Publishing Company.
[12] Bender, E.M., Gebru, T., McMillan-Major, A. and Shmitchell, S. (2021) On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 3-10 March 2021, 610-623. [Google Scholar] [CrossRef] 
[13] Tannen, D. (1993) Framing in Discourse. Oxford University Press.
[14] Goldfarb, D. (2020) Rebooting AI: Building Artificial Intelligence We Can Trust. Computing Reviews, 61, 289.
[15] Creswell, J.W. and Plano Clark, V.L. (2007) Designing and Conducting Mixed Methods Research. Sage Publications.