gpntbНаучные и технические библиотекиScientific and Technical Libraries1027-36892686-8601Russian National Public Library for Science and Technology10.33186/1027-3689-2025-7-142-163gpntb-1558Research ArticleСОВРЕМЕННЫЕ ИНФОРМАЦИОННЫЕ ТЕХНОЛОГИИ. ЦИФРОВАЯ ТРАНСФОРМАЦИЯ ДЕЯТЕЛЬНОСТИ БИБЛИОТЕКMODERN INFORMATION TECHNOLOGIES. DIGITAL TRANSFORMATION OF LIBRARIESМодель коммуникации с искусственным интеллектом ДРУГ как методологический подход к составлению и оценке промптовThe model of communication with artificial intelligence as a methodological approach to prompt creation and evaluationКовалевскийА. В.KovalevskyA. V.

Ковалевский Алексей Викентьевич – магистр педагогических наук, аспирант Белорусского государственного университета культуры и искусств; Научная библиотека Белорусского национального технического университета

Минск

Aleksey V. Kovalevsky – Master of Science (Pedagogy),  postgraduate student, Belarus State University of Culture and Arts, Science Library; Belarus National Technological University

Minsk

kovalevskyalex@yandex.ruБелорусский национальный технический университетБеларусьBelarusian National Technical UniversityBelarus20250308202507142163Copyright © Ковалевский А.В., 20252025Ковалевский А.В.Kovalevsky A.V.Данная работа распространяется под лицензией Creative Commons Attribution 4.0.This work is licensed under a Creative Commons Attribution 4.0 License.https://ntb.gpntb.ru/jour/article/view/1558

Цифровая трансформация библиотечной сферы и постепенное внедрение технологий искусственного интеллекта в практику работы библиотечных специалистов актуализировали проблему эффективного взаимодействия с нейросетями и формирования качественных промптов. В статье представлена авторская модель коммуникации с искусственным интеллектом, основанная на принципах доброжелательности, рациональности, уточнения и гносеологичности (ДРУГ) и разработанная как методологический подход к промптинжинирингу в библиотечной деятельности. Цель исследования – описание данной модели и демонстрация её практического применения и эффективности. Методология исследования включала разработку модели, создание серии промптов различной «силы» на основе её принципов и тестирование с использованием шести моделей нейросетей для решения задачи анализа библиографических данных. Результаты тестирования показали прямую зависимость качества ответов нейросетей от уровня проработки промпта в соответствии с моделью ДРУГ. Практическая значимость исследования заключается в предложении методологической основы для повышения эффективности и качества взаимодействия библиотечных специалистов с искусственным интеллектом, а также в возможности использования модели для разработки практических рекомендаций и образовательных программ по промпт-инжинирингу в библиотечной сфере.

In the context of digital transformation of the library sphere and gradual implementation of artificial intelligence technologies into library practice, the challenge of effective interaction with neural networks and high-quality promptsmithing is increasingly relevant. The author proposes his model of communication with artificial intelligence, based on the principles of Benevolence, Rationality, Refinement, and Epistemology, and demonstrates its practical application and effectiveness. The model is developed as a methodological approach to prompt engineering in library practice. The study methodology comprised model development, engineering of the series of prompts of varying “strength” based on its principles, and testing with 6 neural network models to accomplish the task of bibliographic data analysis. The testing results demonstrate direct correlation between the quality of neural network response and the level of prompt sophistication based on the proposed model. The practical significance of the study lies in offering methodological framework for improving the efficiency and quality of interaction between library specialists and artificial intelligence, as well as the possibility of using the model to develop practical recommendations and educational programs on prompt engineering in the library sphere.

искусственный интеллектбиблиотечное делопромпт-инжинирингпромптнейросетикоммуникацияинформационно-коммуникационные технологиикомпетенциианалитическая работаанализ данныхмоделированиеartificial intelligencelibrarianshipprompt engineeringpromptartificial neural networkscommunicationinformation and communication technologiescompetenciesanalytical workdata analysismodelingReferencesDang H., Mecke L., Lehmann F., Goller S., Buschek D. How to prompt? Opportunities and challenges of zero- and few-shot learning for human-AI interaction in creative applications of generative models // arXiv preprint. 2022. URL: https://arxiv.org/abs/2209.01390 (access: 15.01.2025).Dang H., Mecke L., Lehmann F., Goller S., Buschek D. How to prompt? Opportunities and challenges of zero- and few-shot learning for human-AI interaction in creative applications of generative models // arXiv preprint. 2022. URL: https://arxiv.org/abs/2209.01390 (access: 15.01.2025).Kaddour J., Harris J., Mozes M., Bradley H., Raileanu R., McHardy R. Challenges and Applications of Large Language Models // arXiv preprint. 2023. URL: https://arxiv.org/abs/2307.10169 (access: 14.01.2025).Kaddour J., Harris J., Mozes M., Bradley H., Raileanu R., McHardy R. Challenges and Applications of Large Language Models // arXiv preprint. 2023. URL: https://arxiv.org/abs/2307.10169 (access: 14.01.2025).Zihao Z., Wallace E., Feng S., Klein D., Singh S. Calibrate Before Use: Improving Few-shot Performance of Language Models // Proceedings of the 38th International Conference on Machine Learning, PMLR 139. 2021. URL: https://proceedings.mlr.press/v139/zhao21c.html (access: 21.01.2024).Zihao Z., Wallace E., Feng S., Klein D., Singh S. Calibrate Before Use: Improving Few-shot Performance of Language Models // Proceedings of the 38th International Conference on Machine Learning, PMLR 139. 2021. URL: https://proceedings.mlr.press/v139/zhao21c.html (access: 21.01.2024).Webson A., Pavlick E. Do Prompt-Based Models Really Understand the Meaning of Their Prompts? // Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, 2022. С. 2300–2344.Webson A., Pavlick E. Do Prompt-Based Models Really Understand the Meaning of Their Prompts? // Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Association for Computational Linguistics, 2022. С. 2300–2344.Knoth N., Tolzin A., Janson A., Leimeister J. M. AI literacy and its implications for prompt engineering strategies // Computers and Education: Artificial Intelligence. 2024. Т. 6. URL: https://doi.org/10.1016/j.caeai.2024.100225 (access: 22.01.2025).Knoth N., Tolzin A., Janson A., Leimeister J. M. AI literacy and its implications for prompt engineering strategies // Computers and Education: Artificial Intelligence. 2024. Т. 6. URL: https://doi.org/10.1016/j.caeai.2024.100225 (access: 22.01.2025).Yin Z., Wang H., Horio K., Kawahara D., Sekine S. Should We Respect LLMs? A Cross-Lingual Study on the Influence of Prompt Politeness on LLM Performance // The 2nd Workshop on Social Influence in Conversations (SICon): Proceedings of the Workshop. Association for Computational Linguistics, 2024. С. 9–35.Yin Z., Wang H., Horio K., Kawahara D., Sekine S. Should We Respect LLMs? A Cross-Lingual Study on the Influence of Prompt Politeness on LLM Performance // The 2nd Workshop on Social Influence in Conversations (SICon): Proceedings of the Workshop. Association for Computational Linguistics, 2024. С. 9–35.Пиперски А. Ч. Вежливость в коммуникации между человеком и искусственным интеллектом // Слово.ру: балтийский акцент. Т. 15, № 4. С. 89–98.Piperski A. Ch. Vezhlivost` v kommunikatcii mezhdu chelovekom i iskusstvenny`m intellektom // Slovo.ru: baltii`skii` akcent. T. 15, № 4. S. 89–98.Bass B. M., Waldman D. A., Avolio B. J., Bebb B. J. Transformational Leadership and the Falling Dominoes Effect // Group & Organization Studies. 1987. Т. 12, № 1. С. 73–87.Bass B. M., Waldman D. A., Avolio B. J., Bebb B. J. Transformational Leadership and the Falling Dominoes Effect // Group & Organization Studies. 1987. Т. 12, № 1. С. 73–87.Mayer D. M., Kuenzi M., Greenbaum R., Bardes M., Salvador R. How low does ethical leadership flow? Test of a trickle-down model // Organizational Behavior and Human Decision Processes. 2009. Т. 108, № 1. С. 1–13.Mayer D. M., Kuenzi M., Greenbaum R., Bardes M., Salvador R. How low does ethical leadership flow? Test of a trickle-down model // Organizational Behavior and Human Decision Processes. 2009. Т. 108, № 1. С. 1–13.Foulk T., Woolum A., Erez A. Catching Rudeness Is Like Catching a Cold: The Contagion Effects of Low-Intensity Negative Behaviors // Journal of Applied Psychology. 2016. Т. 101, № 1. С. 50–67.Foulk T., Woolum A., Erez A. Catching Rudeness Is Like Catching a Cold: The Contagion Effects of Low-Intensity Negative Behaviors // Journal of Applied Psychology. 2016. Т. 101, № 1. С. 50–67.Miehling E., Nagireddy M., Sattigeri P., Daly E. M., Piorkowski D., Richards J. T. Language Models in Dialogue: Conversational Maxims for Human-AI Interactions // Findings of the Association for Computational Linguistics: EMNLP. 2024. С. 14420–14437.Miehling E., Nagireddy M., Sattigeri P., Daly E. M., Piorkowski D., Richards J. T. Language Models in Dialogue: Conversational Maxims for Human-AI Interactions // Findings of the Association for Computational Linguistics: EMNLP. 2024. С. 14420–14437.Orme M., Yu Y., Tan Z. How Much do Robots Understand Rudeness? Challenges in Human-Robot Interaction // Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). Torino: ELRA Language Resource Association, 2024. С. 8247–8257.Orme M., Yu Y., Tan Z. How Much do Robots Understand Rudeness? Challenges in Human-Robot Interaction // Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024). Torino: ELRA Language Resource Association, 2024. С. 8247–8257.White J., Fu Q., Hays S., Sandborn M., Olea C., Gilbert H., Schmidt D. C. A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT // arXiv preprint. 2023. URL: https://arXiv:2302.11382 (дата обращения: 14.01.2025).White J., Fu Q., Hays S., Sandborn M., Olea C., Gilbert H., Schmidt D. C. A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT // arXiv preprint. 2023. URL: https://arXiv:2302.11382 (access: 14.01.2025).Prompting Techniques. URL: https://www.promptingguide.ai/techniques (access: 22.01.2025).Prompting Techniques. URL: https://www.promptingguide.ai/techniques (access: 22.01.2025).Li X., Dong G., Jin J., Zhang Y., Zhou Y., Zhu Y., Zhang P., Dou Z. Search-o1: Agentic Search-Enhanced Large Reasoning Models // arXiv preprint. 2025. URL: https://arxiv.org/abs/2501.05366 (access: 22.01.2025).Li X., Dong G., Jin J., Zhang Y., Zhou Y., Zhu Y., Zhang P., Dou Z. Search-o1: Agentic Search-Enhanced Large Reasoning Models // arXiv preprint. 2025. URL: https://arxiv.org/abs/2501.05366 (access: 22.01.2025).Liu R., Geng J., Peterson J. C., Sucholutsky I., Griffiths T. L. Large Language Models Assume People are More Rational than We Really are // arXiv preprint. 2024. URL: https://arxiv.org/abs/2406.17055 (access: 22.01.2025).Liu R., Geng J., Peterson J. C., Sucholutsky I., Griffiths T. L. Large Language Models Assume People are More Rational than We Really are // arXiv preprint. 2024. URL: https://arxiv.org/abs/2406.17055 (access: 22.01.2025).Lewis P., Perez E., Piktus A., Petroni F., Karpukhin V., Goyal N., Küttler H., Lewis M., Yih W., Rocktäschel T., Riedel S., Kiela D. Retrieval-augmented generation for knowledgeintensive NLP tasks // Proceedings of the 34th International Conference on Neural Information Processing Systems (NIPS '20). 2020. С. 9459–9474.Lewis P., Perez E., Piktus A., Petroni F., Karpukhin V., Goyal N., Küttler H., Lewis M., Yih W., Rocktäschel T., Riedel S., Kiela D. Retrieval-augmented generation for knowledgeintensive NLP tasks // Proceedings of the 34th International Conference on Neural Information Processing Systems (NIPS '20). 2020. С. 9459–9474.Jadallah A. K., Mihus I., Svyrydiuk N., Zhyvko Z. Nature and purpose of artificial intelligence. Political, legal, and economic challenges in the 21st century // CLÍO: Revista de Revista de Historia, Ciencias Humanas y pensamiento crític. 2024. № 8. С. 306–320. DOI https://doi.org/10.5281/zenodo.12600350.Jadallah A. K., Mihus I., Svyrydiuk N., Zhyvko Z. Nature and purpose of artificial intelligence. Political, legal, and economic challenges in the 21st century // CLÍO: Revista de Revista de Historia, Ciencias Humanas y pensamiento crític. 2024. № 8. С. 306–320. DOI https://doi.org/10.5281/zenodo.12600350.Bassano C., Caputo F., Barile P., Piciocchi P. Empowering service systems through Intelligence Augmentation (IA) in digital society // ITM Web of Conferences. 2024. Т. 62. DOI https://doi.org/10.1051/itmconf/20246205003.Bassano C., Caputo F., Barile P., Piciocchi P. Empowering service systems through Intelligence Augmentation (IA) in digital society // ITM Web of Conferences. 2024. Т. 62. DOI https://doi.org/10.1051/itmconf/20246205003.Claude. URL: https://claude.ai/ (access: 13.01.2025).Claude. URL: https://claude.ai/ (access: 13.01.2025).ChatGPT. URL: https://chatgpt.com/ (access: 13.01.2025).ChatGPT. URL: https://chatgpt.com/ (access: 13.01.2025).Gemini. URL: https://gemini.google.com/ (access: 13.01.2025).Gemini. URL: https://gemini.google.com/ (access: 13.01.2025).Microsoft Copilot. URL: https://copilot.microsoft.com/ (access: 13.01.2025).Microsoft Copilot. URL: https://copilot.microsoft.com/ (access: 13.01.2025).Mistral AI. URL: https://mistral.ai/ (access: 13.01.2025).Mistral AI. URL: https://mistral.ai/ (access: 13.01.2025).YandexGPT 4. URL: https://ya.ru/ai/gpt-4 (access: 13.01.2025).YandexGPT 4. URL: https://ya.ru/ai/gpt-4 (access: 13.01.2025).Google Colab. URL: https://colab.research.google.com/ (access: 13.01.2025).Google Colab. URL: https://colab.research.google.com/ (access: 13.01.2025).

The authors declare that there are no conflicts of interest present.