Scholarlink

The research aims to evaluate the accurate percentage at which humans can identify text generated by a neural network system (e.g., ChatGPT) within a collection of textual documents used for managerial decision-making. Additionally, the research aims to propose an assessment algorithm that enhances the quality of document recognition, suitable for integration into decision support system frameworks, justify the role of the implemented control (assessment) operation in the process of preparing and executing managerial decisions, and review potential approaches for automating the proposed assessment. The research methodology involves case study analysis examining instances of digital tools used for detecting generated text, document analysis, content analysis employing Osgood’s dependency analysis method, a sociological scientific-practical field experiment, and semantic and syntactic text analysis. Based on experimental data, this research quantitatively measures the actual proportion of managerial documents generated by neural networks during human visual document recognition. Conclusions are drawn regarding the feasibility of using generated documents in the managerial decision-making process. The authors developed a three-tiered assessment algorithm for managerial documents, utilizing quantitative frequency methods (content analysis) and the construction of a C. Osgood matrix, suitable for application in the development and support systems of managerial decision-making at operational and strategic levels.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Applying a Neural Network Algorithm to Detect ChatGPT-Generated Text Within Management Document Repositories as a Tool to Enhance the Efficiency of Decision Support Systems

  • 章节

  • 发表日期：

    2026年

查看全文
  • Sergey V. Bolotnikov,
  • Yaroslav V. Nikulin,
  • Lyubov V. Borodacheva,
  • Lev A. Ivanyan

摘要

The research aims to evaluate the accurate percentage at which humans can identify text generated by a neural network system (e.g., ChatGPT) within a collection of textual documents used for managerial decision-making. Additionally, the research aims to propose an assessment algorithm that enhances the quality of document recognition, suitable for integration into decision support system frameworks, justify the role of the implemented control (assessment) operation in the process of preparing and executing managerial decisions, and review potential approaches for automating the proposed assessment. The research methodology involves case study analysis examining instances of digital tools used for detecting generated text, document analysis, content analysis employing Osgood’s dependency analysis method, a sociological scientific-practical field experiment, and semantic and syntactic text analysis. Based on experimental data, this research quantitatively measures the actual proportion of managerial documents generated by neural networks during human visual document recognition. Conclusions are drawn regarding the feasibility of using generated documents in the managerial decision-making process. The authors developed a three-tiered assessment algorithm for managerial documents, utilizing quantitative frequency methods (content analysis) and the construction of a C. Osgood matrix, suitable for application in the development and support systems of managerial decision-making at operational and strategic levels.