Quote of the Week: "These tasks can no longer be used"

Content:

• Context

Context

In a column for RBC+, Ivan Karlov expressed his opinion on the influence of generative artificial intelligence (GAI) technologies on traditional pedagogical practices. GAIs are capable of generating text, images, videos, and other content in response to simple queries, challenging established teaching methods. Karlov emphasizes the need to adapt educational strategies to effectively integrate new technologies and maintain the quality of education. It's important to consider how GAIs can complement, rather than replace, the educational process, opening up new opportunities for learning and development.

The problem identified after the launch of the ChatGPT chatbot has become noticeable primarily in the United States. It concerns educational practices, particularly in the humanities. Traditionally, the educational process is focused on text-based assignments, including essays, papers, and term papers. Graduate theses also require a significant amount of written text. However, with the advent of neural networks, students can now generate these works independently, and the results are often comparable to high-quality human work. This raises serious questions about the effectiveness of traditional assessment methods and requires a reconsideration of approaches to teaching and assessing students in the face of technological change.

Ivan Karlov noted that it was previously believed that students' essay-writing and analytical skills were demonstrated in texts submitted to instructors. However, using generative AI tools, we can only assess how effectively a neural network performs a given task. This raises important questions about the integrity of assessment and how technology is changing approaches to teaching and assessing student knowledge.

Problems with academic dishonesty existed even before the advent of generative artificial intelligence (GAI) technologies. It's no secret that students often cheated or purchased ready-made assignments. However, GAI seems to have finally eliminated the possibility of completing assignments that facilitate cheating. As Evgeny Patarakin, professor in the Department of Informatics, Management, and Technology at Moscow State Pedagogical Univ. and leading expert at the HSE Laboratory for Digital Transformation of Education, noted, "nothing has changed in terms of student cheating itself. If it's advantageous for a student to present someone else's work as their own, then it makes no difference how they do it—buying the work, hiring someone else, or using ChatGPT. Therefore, it's necessary to restructure the educational process so that this becomes disadvantageous." Optimizing the educational process and using innovative technologies should be a priority to improve academic integrity.

Changes in the organization of work are necessary, and this applies to several aspects. First and foremost, it's important to optimize internal processes to increase efficiency. This may include reviewing current working methods, implementing new technologies, and training employees. It is also worth paying attention to interteam collaboration and improved communication, which will help avoid duplication of tasks and increase productivity. It is important to consider employee feedback to identify bottlenecks and find solutions. Optimizing work at several levels will help achieve better results and increase competitiveness.

• Faculty need to develop assignments that students cannot simply copy the answers to from dialogues with chatbots.
• Practices are already spreading at the university level, where the use of generative artificial intelligence in the preparation of coursework and theses is not only not prohibited, but even encouraged (prohibiting it is pointless, since students will still circumvent these restrictions).

This year, the Higher School of Economics organized a competition for the best theses completed using neural networks among final-year students. The university allows the use of generative artificial intelligence (GAI), however, when submitting a paper, it is necessary to indicate which tasks were transferred to neural networks and the rationale for this choice. HSE University is not the only Russian university to support the use of generative neural networks in the educational process. Previously, Moscow State University of Psychology and Education (MSPU) and Northern Arctic Federal University (NArFU) announced permission to use such technologies in academic writing. This demonstrates the growing interest in integrating new technologies into the educational process and the relevance of exploring their potential.

Read also:

Quote of the week: it is necessary to change the role of a knowledge relay to a "cognitive engineer". This transformation emphasizes the importance of the active participation of specialists in the process of processing and interpreting information. A cognitive engineer doesn't simply impart knowledge; they create new approaches to learning and development, applying modern methods and technologies to optimize information comprehension. With rapidly changing technologies and increasing data volumes, this role is especially relevant, as it facilitates deeper understanding and the effective use of knowledge in various fields.

There are several reasons why changing educational practices is challenging. Ivan Karlov highlighted the main ones.

• Teachers require additional training to work in this new environment. Incidentally, according to a recent survey conducted by researchers from RANEPA, only 16% of Russian university professors currently regularly use AI in their work. Only among those teaching computer science is this proportion significantly higher—34%.
• The transition to new types of assignments means that students will be using skills for which there are no proven assessment methods yet. Ivan Karlov gives the example of preparing a thesis: now there's no need to waste time writing the text—instead, the student must quickly and correctly compose prompts for the neural network, edit, and verify its responses. It turns out that these skills, not the text itself, need to be assessed. However, there is no established and replicable practice for assessing the quality of work with AI tools in education, and this is a problem.

Overcoming the existing problems in the education system remains an open question. Diana Koroleva, head of the Laboratory for Innovation in Education at the HSE Institute of Education, emphasized that over the past 20 years, the system hasn't adapted to tools like Google. Now, a new challenge has emerged: ChatGPT, which can not only provide straightforward answers but also solve complex problems. Mikhail Kushnir, a researcher at the Institute of Social Sciences at RANEPA and a board member of the Education League, highlighted the general rigidity of the education system. It adapts slowly and with difficulty to new approaches, which slows the implementation of modern technologies and teaching methods. The need for educational reform is becoming increasingly urgent, especially in light of the rapid development of technologies that impact learning and interaction with information. The emerging problems cannot be ignored. The phenomenon of generative artificial intelligence calls into question the relevance of the traditional ideas on which the education system is still based. This opinion was expressed by Pavel Sorokin, head of the Laboratory for Human Potential and Education Research at the Institute of Education at the Higher School of Economics, a year ago. Given the rapid development of technology, it is necessary to rethink the organization of the educational system as a whole.