Contents:
Why AI is a disruptive technology for universities
The Center for Educational Development Based on AI Technologies at Tyumen State University (TSU) was recently created, which became an important step in the development of the educational process. The center's primary goal is to integrate modern technologies into the educational process, thereby improving the quality of education and preparing students for the demands of the modern labor market. The center also aims to develop innovative educational programs that use artificial intelligence to tailor learning materials to the individual needs of students. This significantly improves the effectiveness of learning and prepares specialists with relevant knowledge and skills in IT and related disciplines.
The center was recently founded, after the first series of experiments with artificial intelligence were completed.
Artificial intelligence is viewed as a revolutionary technology capable of transforming the traditional educational process at universities. When students use AI for plagiarism, it undermines the entire educational process, hinders the achievement of educational goals, and deprives it of meaning. Banning the use of AI is not an effective solution, as graduates in the future will be faced with the need to apply new technologies in their professional activities. Universities must respond to changes in the world of technology and integrate them into the educational process, not ignore them. This will prepare students for real-world challenges and ensure their competitiveness in the labor market.
The use of artificial intelligence at the university must be justified and targeted. Our center focuses on radical tasks aimed at experimentally testing hypotheses about the use of AI in the educational process. We are also developing innovative educational models that can significantly improve the quality of education.
At what point did you realize that the influence of artificial intelligence technologies was beginning to have a disruptive impact?
In the fall of 2023, it became evident that students at the School of Advanced Studies were actively using artificial intelligence to write essays, which are one of the key forms of educational activity. We noticed that the quality of work submitted by students at the end of their courses increased significantly compared to their previous results in the classroom. This indicates that traditional assessment methods are becoming less effective and require revision.
Director of the School of Advanced Studies Andrey Shcherbenok notes that previously, analysis of final theses made it possible to track a student's development and progress from their first year. However, as artificial intelligence becomes an increasingly common tool for writing these papers, the reliability of this analysis is diminishing. Theses generated by AI do not reflect students' true knowledge and skills, calling into question their value as a measure of educational growth. It is important to consider these changes to maintain the quality of education and the assessment of student achievement. While oral examinations remain relevant with the advent of artificial intelligence, their effectiveness in assessing educational outcomes raises many questions. This underscores the need to rethink the educational process to take modern technologies into account. It is also important to determine the specific outcomes that universities and educational programs should focus on, a more complex and multifaceted issue. You are considering redesigning the educational process based on the results of your experiments. One key aspect of this model is the transfer of some teacher functions to artificial intelligence. This approach could significantly transform traditional teaching methods, improving their effectiveness and accessibility. The use of artificial intelligence in education offers new opportunities for personalizing learning, automating assessments, and providing feedback to students.
In spring 2024, we began a series of experiments aimed at optimizing the educational process using artificial intelligence. Teachers play a key role in learning, and before beginning these experiments, we carefully analyzed their tasks to determine which could be delegated to AI. This will allow teachers to focus on more complex aspects of learning and improve the quality of education.
We identified five key teacher functions and examined each in detail, breaking them down into specific tasks. These functions are broad, and analyzing individual tasks allows us to assess how effectively teachers perform them compared to artificial intelligence. This approach helps make informed decisions about the feasibility of transferring certain tasks to digital services.
You have chosen a radical A decision to replace a faculty member instead of considering various supporting features of artificial intelligence. Why did you make this decision? Universities, like many other organizations, face the challenge of a shortage of qualified personnel. There is often a significant shortage of faculty with the necessary expertise. While it is possible to fill vacancies, finding specialists who meet internal requirements can be difficult. Furthermore, there is the problem of inefficient faculty workload: significant time is spent on routine tasks, which limits opportunities for complex and time-consuming projects, as well as for personal and professional development. It is important to ensure students have access to highly qualified faculty. Therefore, we began implementing artificial intelligence technologies, with an emphasis on resolving personnel issues and optimizing the educational process. A teacher performs several key functions that play a significant role in the educational process. First, they are a source of knowledge, conveying information to students and explaining complex topics. Second, the teacher acts as a mentor, supporting and guiding students throughout the learning process. Another important function is performance assessment, which allows us to determine the level of mastery of the material. In addition, the teacher creates an atmosphere for discussion and exchange of opinions, which promotes the development of critical thinking in students. Each of these functions contributes to the formation of a high-quality educational experience and helps students achieve their goals.
The first function of the teacher is the provision of expert knowledge. This role is traditionally associated with teaching, when the teacher comes to the classroom to share their knowledge and experience.
The second role of the teacher is the cartographer of the subject area. In this role, they help students understand the structure of the subject area, identifying already resolved questions and those that remain open. The teacher also highlights current debates occurring in the field and demonstrates how it interacts with other disciplines. This promotes a deeper understanding of the material and forms a holistic view of the subject matter.
The third key function of the teacher is the organization of learning activities. It is important to understand that educational outcomes such as knowledge, skills, thinking style, and attitudes cannot simply be transferred from the teacher to students. It is necessary to create conditions for active student activity, through which these outcomes will be formed. Within this function, the teacher faces a variety of tasks at various stages of the work—from designing learning activities to analyzing their results. Effective organization of the learning process facilitates deeper learning and the development of essential competencies in students.
The fourth function—developing interpersonal experiences—emphasizes the importance of human contact in education and learning. Effective interaction between teacher and student, as well as between students themselves, plays a key role in creating a productive learning environment. Such contact promotes better learning, fosters collaboration skills, and develops social competencies.
An interesting observation is that mimetic desire significantly influences educational outcomes. This is students' desire to experience the same emotions and states as the teacher, for example, while they are explaining a subject area. In this context, the teacher becomes a role model, which can facilitate deeper learning. Such interactions between teacher and students create an atmosphere in which students strive not only to understand the subject but also to share their mentor's enthusiasm and interest.
The fifth function is the keeper of motivation, which plays a key role in the learning process. Not all students are initially highly motivated, and as their studies progress, interest inevitably declines. Even with initial enthusiasm, it quickly becomes apparent that successfully mastering a new discipline requires studying a significant amount of information, understanding numerous issues, learning to communicate in a new language, and applying complex approaches. At this stage, students may encounter a motivational pit, from which they gradually emerge as their competence improves. Maintaining motivation in learning helps students overcome difficulties and maintain a desire for knowledge, which ultimately contributes to their success in their studies and professional activities.
Reworked text for SEO:
Learn Also:
With the digitalization of education, the role of the teacher is undergoing significant changes. Modern teachers are becoming not only a source of knowledge but also a role model for their students. In the age of information technology, teachers must be able to adapt to new realities, using digital tools to improve the effectiveness of learning.
Teaching requires teachers not only professional knowledge but also skills in working with modern technologies. This includes the ability to create interactive educational materials, conduct remote classes, and maintain student interest in the learning process. It is important for teachers to demonstrate critical thinking and digital literacy skills, inspiring students to follow their example.
An important task for modern teachers is to develop self-education skills and responsibility for their own learning in students. Given the rapid development of technology, teachers must motivate students to use digital resources to broaden their horizons and deepen their knowledge.
Thus, in the context of the digitalization of education, the teacher becomes not only a conduit of knowledge, but also a mentor, capable of guiding students on the path of independent and critical thinking in the world of information.
How the experiments were conducted
During our experiments, AI replaced the teacher in several key functions. The main ones are adapting educational material to the individual needs of students, automating assessment of assignments, providing feedback, creating interactive learning tasks, and real-time support. These functions allow AI to effectively interact with students, providing a personalized approach to learning and improving the quality of education.
In the experiments we conducted with various courses, artificial intelligence often played the role of a source of expert knowledge and a cartographer of the subject area. In turn, the other three functions were performed by a person, whom we call a mediator. This term differs from the traditional concepts of "teacher" and "professor" because it highlights an important distinction: a mediator is not an expert in the course topic and, therefore, cannot act as a source of expertise or a cartographer of the subject area. Artificial intelligence also partially takes over tasks related to the organization of the educational process, for example, by providing students with feedback according to specified criteria. This interaction between humans and artificial intelligence opens new horizons in educational technologies and improves the effectiveness of learning.
It turns out that such important "human" functions as configuring interpersonal experiences and maintaining motivation were not transferred to artificial intelligence in experiments. This raises questions about the ability of AI to effectively interact with people and support their emotional state. Configuring interpersonal experiences and maintaining motivation remain key aspects that require a deep understanding of human nature and emotional intelligence, which is not yet available with current technologies.
We are actively discussing the possibility of transferring certain functions to artificial intelligence. Students may not distinguish whether they are communicating with a person or a bot. It is important to understand how such communication can impact educational outcomes. After the first series of experiments on other functions, many questions remain that require further study.
The functions you mentioned were performed by the artificial intelligence systems with which students interacted during the experiments. Please describe the process of creating these AI bots and their characteristics.
Our "AI personas," as we call them, are developed based on existing neural networks, as we do not create our own models. These are customized solutions with unique capabilities specifically designed for the educational process. For example, they can provide feedback in a specific manner and also use pre-selected texts and information sources. For students, these AI personas function as Telegram bots, making interaction with them convenient and intuitive. In several courses, AI personas reconstruct the personalities of famous experts in various fields, such as Robert Sapolsky, Sergey Semenovich Uvarov, and Steve Jobs. This allows students to immerse themselves in the subject matter through the experience and knowledge of outstanding individuals, significantly enriching the learning process.
The preparatory work on developing chatbots before launching the courses took a significant amount of time. This stage includes analyzing the needs of the target audience, designing the functionality and interface, and testing the bots. All of these processes are necessary to ensure the chatbots operate effectively and achieve their goals. High-quality preparation not only improves user interaction but also increases overall course satisfaction.
Developing educational courses can vary in complexity, but on average takes one to two weeks. This timeframe turned out to be significantly shorter than initially expected. The process does require significant effort, including the active participation of engineers and the input of professors or course facilitators. Professors must select the texts on which the bot is trained and develop a consistent workflow for students. This includes clear steps that students must follow to achieve the desired results. This approach ensures structured and effective learning, which in turn improves the quality of the educational process.
Creating such tools takes time, but the amount of work is not so great that we should abandon this process.
We carefully selected the courses where chatbots would perform teaching functions based on several key criteria. First, we analyzed the course content to ensure that the topics were suitable for automation and could be effectively presented in a dialog format. Second, we conducted an interaction test with potential users to determine how well the chatbot could meet their educational needs. We also considered the possibility of integrating chatbots with existing educational platforms and learning management systems. Finally, we evaluated the technical characteristics of the chatbots, including their ability to adapt to different learning styles and provide a personalized approach to each student. These aspects helped us select the courses in which chatbots would be most effective as teachers.
In one case, we were unable to find a teacher for a new course. This is due to the fact that the discipline is just beginning to develop, and its creation was initiated by senior students. As a result, we were faced with the need to find an alternative solution to replace the non-existent teacher.
We selected some courses based on the preferences of faculty willing to experiment. One such course was design thinking, which I represent. Additionally, we sought to include a variety of disciplines: courses with clear algorithms on the one hand, and those that require the faculty to actively engage with students' thinking on the other. This ensures the comprehensive development of students and creates a diverse educational environment.
It is important to explore the requirements for technological solutions and their impact on students using various examples. The decision to include certain courses in the experiment was made by the director without consulting with faculty or students. This approach also represents an important experience in the context of the educational process.
When researching generative artificial intelligence, we had certain unshakable postulates. We proceeded from the belief that AI is capable of creating unique content that can be useful in various fields. We also believed that machine learning algorithms should be based on ethical principles that ensure safety and transparency. These ideas formed the foundation of our research approach, and we did not plan to revise them, as we considered them key to understanding the capabilities and limitations of generative AI.
One of the main challenges of generative AI is its tendency to "hallucinations," that is, to misrepresent unreliable information as fact. In everyday life, this problem can be overcome by double-checking the answers received. However, in education, such a situation is unacceptable, as students should not be exposed to false information. While the ability to critically perceive information can be beneficial, the constant need to question the credibility of sources makes the educational process ineffective. We proceeded from a key principle: our chatbots must provide students with only verified information. All other aspects of the work required further verification and optimization.
To achieve optimal results in SEO, it is important to create high-quality content that is not only informative but also attractive to users. Text optimization involves using topic-related keywords and structuring information for easier comprehension. Ensure your text answers your audience's questions and provides valuable information. This will help improve search engine visibility and attract your target audience. Don't forget to regularly update your content to keep it relevant and current with modern user needs. Read also:
AI assistants in education have acquired a wide range of capabilities that significantly improve the learning process. They can adapt educational materials to individual student needs, providing personalized recommendations and assignments. Such technologies help automate routine tasks, such as checking homework and grading tests, allowing teachers to focus more time on teaching.
Furthermore, AI assistants can analyze student progress, identify their strengths and weaknesses, and suggest additional resources to improve their knowledge. Using chatbots in teaching allows students to get answers to their questions at any time, which contributes to a deeper understanding of the material.
It is also worth noting that AI assistants are becoming an important tool for increasing student motivation, offering interactive and engaging learning experiences. The introduction of these technologies into the educational process opens new horizons for students and teachers, improving the quality of education and making it more accessible.
Results
Several key findings were reached based on the first series of experiments. First, the results showed that the methodology used in the research was effective and yielded reliable data. Second, the identified patterns confirm the hypotheses put forward at the initial stage. This opens new opportunities for further research and development in this area. Finally, the obtained results can serve as a basis for practical application and improvement of existing processes.
A useful result is that instructors and course facilitators have begun to rethink the educational process. This demonstrates a desire to improve the quality of learning and adapt teaching methods to modern requirements. Rethinking teaching approaches allows for more effective meeting student needs and ensuring a deeper understanding of the material.
Modern universities still widely use the broadcast teaching model, in which the instructor simply conveys their knowledge in a monologue. In this situation, little attention is paid to how students perceive information. For an hour and a half of class, they may spend time passively listening, without engaging in active learning. It is assumed that outside of the classroom, the student will independently understand the material, learn, and read the necessary literature. However, even the introduction of artificial intelligence into this model will not lead to positive results, as both instructors and students may ignore its capabilities. To improve the effectiveness of teaching, it is necessary to reconsider approaches and implement more active and interactive methods that promote better assimilation of material and student engagement in the educational process.
Replacing the teacher with artificial intelligence in the educational process raises the question of their role. What should the teacher do if the functions of retelling information are no longer needed? This requires a radical redesign of the educational model, as well as the design of both student activities and their own. The teacher must focus on achieving specific educational outcomes, using new approaches and methodologies that meet modern requirements.
The educational process is transforming from a teacher-centered approach to a student-centered one. It is important to focus on what students will do within the course. The emphasis shifts to active participation, where students not only listen to lectures but are also involved in practical activities. This allows for the development of their skills and competencies, which is a key aspect of modern education.
Of course, we can consider an example of changes in a specific course. For example, the web development course was updated with a new curriculum that incorporates current technologies and tools. Outdated teaching methods have been replaced with a focus on practical assignments, allowing students to quickly grasp the material and apply it to real-world projects. This course update has significantly increased its popularity and attracted more students interested in gaining up-to-date knowledge in web development.
In the "World Through Time" course, students explore the analysis of historical events from various perspectives, including cultural, political, and economic. Before the introduction of AI characters, the course consisted of a traditional lecture series, where the instructor shared their personal approach to analyzing historical events. Now, with the use of AI characters, the learning process has become more interactive, allowing students to independently explore and discuss various aspects of history, fostering a deeper understanding of the material and developing analytical skills.
Artificial intelligence plays an important role in the educational process, but its use must be carefully considered. To increase student engagement, a change in the approach to teaching is necessary. Instead of simply listening to lectures on analytical methods from experts, students must actively participate in the process by conducting their own analysis. Artificial intelligence in this context becomes a valuable tool, providing access to expert knowledge and assisting in solving practical problems. This not only promotes a better understanding of the material but also develops the skills necessary for successful professional work in the future.
Analyzing historical events requires preparation, and we have developed a program that takes this into account. Before key events, students undergo a preparatory phase. To facilitate the comprehension of information, we use chatbots based on the texts of historical figures. This allows students to interact with key historical figures: ask questions, express doubts, and receive subjective opinions. This approach promotes a deep understanding of events and develops critical thinking, which is especially important in the study of history.
Students interacted with AI bots that imitated individuals involved in significant historical events. This allowed them to deepen their understanding of the historical context and develop critical thinking skills through dialogue with virtual characters. Using such technologies in the educational process opens new horizons for the study of history, making it more interactive and engaging.
A chatbot trained on Stalin's texts is capable of responding in the spirit of the historical figure, which I believe helps students develop a deeper personal connection to important historical events. Using such technologies in the educational process not only enlivens the study of history but also creates an interactive environment in which students can better understand the context and significance of events.
Students demonstrated high effectiveness in working with AI personas in this and other courses. Using artificial intelligence in the educational process allowed for a deeper understanding of the material and increased interaction. AI personas became a valuable tool for modeling various scenarios, which contributed to the development of critical thinking and practical skills in students.
Students demonstrated a low level of interaction with artificial intelligence, which surprised us. Although they engage with the AI, the frequency of such interactions is significantly lower than expected. We also noticed that AI is not perceived as a source of authoritative information: answers received from AI have much less value for students compared to the opinion of a professor or other high-status person. Questions generated by AI often go unnoticed by students.
Read also:
The question of whether people will be interested in learning created by artificial intelligence is becoming increasingly relevant. Artificial intelligence can analyze vast amounts of data and tailor the learning process to the needs of each student, making such learning more personalized and effective. However, it's important to keep in mind that the success of AI-based learning depends on the quality of the content and its delivery methods. People may be interested in such programs if they provide a high level of engagement and practical value. Therefore, to increase the appeal of learning, developers need to focus on creating interactive and accessible content that meets the current needs and interests of learners.
Another important finding: students don't know how to ask questions. In the modern educational system, this is often not a necessary requirement. If students don't take the initiative to ask questions, this doesn't cause concern for teachers. However, the ability to ask questions is a key skill that promotes deep understanding and the development of critical thinking. Fostering curiosity and active participation in the learning process can significantly improve the quality of education and prepare students for future challenges.
I want to share an illustrative story. A friend's son was accepted to university after being homeschooled. He had never been in a classroom before, only working with a tutor. After his first few days of school, he confided in his mother that he felt like everyone else understood everything, while he alone was struggling to grasp the material. She, naturally, was alarmed and asked why. He replied, "Everyone's quiet, except me." This situation highlights the importance of open communication and active participation in the learning process. Students need to understand that asking questions is normal and necessary for successful learning. Students don't expect to be required to ask questions, which leads to a lack of skill in formulating questions to gain a deep understanding of the topic. This lack of engagement in learning impacts their analytical and critical thinking abilities. Without questions, it is impossible to fully grasp the material, which negatively impacts the learning process and skill development. It's important to create an environment where students are motivated to ask questions and actively participate in discussions, which will help them gain a deeper understanding of the topics they're learning. Yes, that's true. During our interview, we discussed how the education system doesn't teach children, adolescents, and adults how to ask questions. This leads to difficulties when interacting with artificial intelligence. A lack of question-formulating skills limits the ability to effectively use AI. It's important to develop the ability to ask relevant and insightful questions to maximize the potential of technology and receive valuable answers. Artificial intelligence (AI) currently answers our questions, but it doesn't take the initiative. This creates challenges for students who haven't been trained to ask questions to gain a deeper understanding of the material, but rather to simply listen. Unlike AI, a human teacher can adapt their responses to students' needs, interpreting unsuccessful questions and suggesting more relevant ones. Teachers can guide students to important aspects of a topic, which helps them learn. In contrast, AI follows strictly defined algorithms and doesn't intuitively understand which questions to ask. We are currently working to train AI to deliver a more flexible and interactive approach to learning, which will help improve the quality of the educational process.
In the second wave of our experiments, we plan to delve into new aspects of the research that were previously beyond our attention. We will focus on analyzing the data obtained in the first wave to identify patterns and identify areas for further study. We also intend to implement new methodologies and technologies that will improve the accuracy and efficiency of our experiments. We expect the results of the second wave to provide a deeper understanding of the topic and open new horizons for research.
The second wave of experiments has already begun, and as part of this phase, we plan to test a number of hypotheses that were not tested in the first wave. The initial experiments were "boutique" in nature, as they involved a limited number of students, and it was not always possible to create control groups. Now we are scaling up, applying the same ideas to a larger number of students. For example, our experiments involve courses in mathematics and Russian history, which are taken by two thousand students at Tyumen State University. These subjects are common at most universities, allowing the results to be replicated and applied at other educational institutions.
Artificial intelligence personas in these massive courses will perform various tasks aimed at improving the learning process. They will be able to adapt content to individual student needs, provide personalized recommendations, and assist in mastering the material. Furthermore, these AI personas will analyze student performance, identifying weaknesses and suggesting additional resources to address them. They can also provide feedback, answering student questions in real time. Thus, integrating AI into massive courses can significantly improve the quality of education and make it more accessible.
The mathematics course has a knowledge graph, which is a map of all the topics students must cover during the semester. A special AI bot analyzes each student's level of mastery of the material and suggests the optimal trajectory for further learning. This ensures the complete acquisition of the necessary skills and knowledge. The bot is designed for independent study, while a traditional teacher continues to conduct classes in the classroom.
Have you considered inviting your colleagues from other universities to conduct similar experiments with AI personalities? This could facilitate the exchange of experience and expand research horizons in this area.
We are actively developing R&D projects in the field of "Artificial Intelligence in Higher Education." Each experiment requires significant effort, but at the same time, it provides important answers to specific questions. To achieve significant results, it is necessary to conduct numerous experiments and involve a large team of specialists. Therefore, we are open to collaboration with researchers from various universities. Moscow City Pedagogical University, Tomsk State University, and Southern Federal University have already joined the project, and we expect the participation of other colleagues.
Some courses developed in the first wave of experiments have already become full-fledged AI products, and we are ready to offer them to other educational institutions. For example, the course on design thinking has already found its external customer.