Использование генеративных моделей в высшем гуманитарном образовании: опыт Института медиа НИУ ВШЭ

Коммуникации. Медиа. Дизайн. 2025. Т. 10. № 1. С. 5–37.

Sharikov A., Dzhura A., Magera T., Onegov D., Khasavov A.

The article delves into the theoretical foundations and practical applications of generative models in higher humanitarian education in Russia. It explores some ethical concerns surrounding the use of generative artificial intelligence. The paper describes how HSE University has been setting up a regulatory framework for its use since 2023. This includes legalizing the use of generative models for educational purposes. One of the key findings of the article is a study conducted among graduate students at the Institute of Media, Faculty of Creative Industries, HSE University. The study aimed to investigate the students' experiences with neural networks. A survey of students from four master's degree programs (n=98) showed that most of them had worked with neural networks before their legalization at the university. It also revealed that students most often used neural networks not to generate texts but to solve problems related to systematizing and structuring textual data. Some students refused to use neural networks. Their position was expressed by the phrase "This devalues my own work, I prefer to do tasks at the university on my own." This attitude was also linked to a general distrust of neural networks. Based on the conducted research, methodological provisions were formulated expressing a rational attitude towards the use of neural networks by students, namely: the interpretation of the generative model as a tool for solving a certain class of problems; understanding the boundaries of the applicability of a neural network: understanding the limited resources when using it; knowledge of the general scheme (algorithm) of its application; the need to verify the result. The article concludes with a positive example of how a generative model was used in the final qualifying work of a student in the master's program "Media Management" by B.A. Podrozhantseva.

Keywords: искусственный интеллект, нейросеть, adaptive neural networks, генеративная модель, generative model