Intelligent Educational Environments: Recent Trends, Modeling, and Applications

Abstract

This paper aims to summarize recent trends in the field of intelligent education and build on the research findings to propose a conceptual model of an intelligent educational environment (IEE) along with possible applications. A bibliometric analysis explores the recent trends to obtain insights into the current research directions. The concept of an IEE is introduced and correlated to other popular concepts in the educational field. Relevant user requirements for the IEE are defined and discussed. They serve as the basis for creating the conceptual model as a UML class diagram. The proposed model gives an overall picture of the relationships among the essential elements of the IEE and provides insights into its future development. Two possible applications are outlined: personalized teaching and learning and the assessment process. Any particular application of the model to build an IEE must consider all specific user requirements and add relevant details. The proposed conceptual model can be used to enhance the teaching–learning process and improve learners’ performance.

1. Introduction

The recent advancement of information and communication technologies (ICTs) and intelligent technologies, especially, provides conditions for the emergence of a new paradigm—intelligent or smart education. Usually, this innovative form of education occurs in a technology-rich, intelligent environment that allows more engaging, adaptive, and personalized pedagogical practices. Thus, such an intelligent educational system (IES) provides opportunities for a more efficient teaching–learning process.

Uskov et al. are among the first to highlight the role of ICT in facilitating interactive education [1]. The concept of IES extends the smart classroom model that integrates many educational technology components with intelligent technologies such as learning analytics (LA), artificial intelligence (AI), ubiquitous, and cloud computing to provide an enhanced learning experience [2]. Further, much research extends the contemporary understanding of IES, where it involves the use of a wide range of modern educational technologies, including networked interactive devices [3,4], the Internet of Things (IoT) [5,6], software applications for data gathering, monitoring, processing, and management [7,8].

Utilizing various intelligent technologies, such as learning analysis, big data, knowledge representation and management technologies, etc., in the teaching process is considered a way to transform an educational environment into a smart one [9,10]. All such technologies enable the collection, processing, and analysis of data related to the teaching–learning process to gain insight into it, thus making amendments necessary to provide an enhanced learning experience [11]. One of the significant advantages of IES is the opportunity for personalization and adaptation of the teaching–learning process [12]. Personalized learning is usually based on students’ profiles, including their knowledge levels, learning goals, interests, etc. Neural networks may be applied to model the changing levels of students’ knowledge at each learning step [13] or to create a mathematical model for students’ categorization using machine learning techniques [14]. Ontologies and clustering methods also can be used to develop student profiles [15]. Intelligent Tutoring Systems (ITSs) [16] can be considered as an early predecessor of IES.

Some researchers [17] also stress the potential impact of the environmental parameters of smart classrooms on the learning process, thus emphasizing the interrelations between physical surroundings and educational outcomes. In this context, the concept of an intelligent educational environment (IEE) can be considered. The IEE controls both the intelligent educational system and the physical environment in which the intelligent education process occurs in an integrated manner. Hence, IEE provides smart management of parameters of physical space (temperature, air quality, light, noise, etc.) and data from external sensors to provide optimal conditions and support the intelligent educational process.

In the wider context of intelligent education, the concepts of IEE, IES, and ITS have distinct functionalities while being related as “a whole” and “parts of the whole”. Such a structured approach considers IEE to be the broadest concept; thus, IES and ITS are parts of it [7]. Both IES and ITS have the same goal: to enhance the learning process utilizing technology, particularly AI and data-driven approaches [9,16]. ITSs have many contributions to the personalization of the learning and teaching process that can be successfully used in IEEs. An ITS can be a core component of an IEE, enabling personalized learning experiences through artificial intelligence and knowledge representation technologies.

The research goal of the current paper is to explore recent advancements in the field of intelligent educational environments and, based on them, to propose a new conceptual model of an intelligent educational environment and its possible applications concerning future implementation. For this reason, the research seeks answers to the following research questions (RQs):

RQ1: 

What are recent topics and research trends regarding intelligent education?

RQ2: 

What are the main user requirements for intelligent educational environments?

RQ3: 

What essential elements and functionalities should be integrated into the conceptual model of an intelligent educational environment?

RQ4: 

What are the possible practical applications of the proposed conceptual model of an intelligent educational environment?

The paper is structured as follows. In Section 2, the methodology used is described. An overview and brief analysis of the recent trends are presented in Section 3. Section 4 describes the concept of an intelligent educational environment. The section considers the related contemporary research on intelligent educational systems and intelligent tutoring systems. Section 5 defines and discusses the user requirements for the IEE, which serve as the basis for creating the corresponding conceptual model. Some possible applications of the model are outlined. The last two sections provide a brief analysis of the studies performed so far. The authors’ views on the answers to the research questions are presented and discussed, and some specific future research directions are outlined.

2. Materials and Methods

The methodology used for achieving the research goal and answering the research questions includes several procedures:

• A bibliometric analysis is performed to outline the overall global picture of research focused on intelligent education. The bibliometric data are taken from the scientific database Scopus upon request for “intelligent AND education” in titles, abstracts, and authors’ keywords.
• To obtain more focused insight, a second query, “intelligent AND education AND environment,” is submitted to the Scopus database, and the results are analyzed.
• A review of implemented intelligent educational environments is conducted to clarify the technologies used, their advantages, and their benefits for individual users in the learning process: teachers and students. Based on the review, users’ requirements for IEE are derived.
• A Unified Modeling Language (UML) model is proposed that shows important elements of an intelligent educational environment, followed by their detailed description.
• Several applications are presented to practically demonstrate the functionalities of intelligent educational environments—the workflow of personalized learning and assessment.

The presented methodology covers the defined research goal and helps answer the RQs. It includes a bibliometric study, thematic analysis, requirements formulation, system modeling, and outlining practical applications of Intelligent Educational Systems.

3. Overall Picture Regarding the Intelligent Education

Over the past ten years (2014–2024), the topic of intelligent education has been very much researched and discussed, as evidenced by the large number of published scientific articles. The query “intelligent AND education” was submitted on 14 March 2025 to the scientific database Scopus. It aims to outline the trends in the research on intelligent education. The query is broad but has the advantage of giving an overall picture of the studies and pointing out trending topics. The goal is to explore the most popular research topics instead of making a comprehensive literature review.

The query is conducted on the titles, abstracts, and author keywords of the publications. The Scopus scientific database has a relatively wide interdisciplinary scope and indexes various research publications worldwide, including conference proceedings, journal articles, and books. That is the main reason the authors conducted the bibliometric query there. The bibliometric study gives valuable insights into current subjects of interest in intelligent education. It is not a thorough literature review on intelligent education but rather a snapshot of the research interest in the explored area.

The returned result contains 15,718 documents. Among these, there could be some out-of-scope publications; however, the presumption is that mentioning the words “intelligent AND education” in the title, abstract, and author keywords provides a solid basis to show the research topics in the area. The extracted bibliometric information of these publications is ordered by relevance and analyzed using RStudio 2024.12.1 and the bibliometrix:biblioshiny application. The annual growth rate is 10.68%, noticing a trend of a continuously increasing number of publications, as in 2014, there were 941; in 2020, there were 1486 (about 1.58 times more than in 2014); and in 2024, there were 2596 (about 2.76 times more than in 2014) (Figure 1). This can be explained by the development of intelligent technologies and their impact on education. Researchers observe how these technologies can contribute to enhancing learning and teaching and improve the quality of education in general.

Table 1. Main information considering the results from the query “intelligent AND education”.

Parameters	Query “Intelligent AND Education”
Timespan	2014–2024
Sources (journals, books, etc.)	4993
Documents	15,718
Annual growth rate %	10.68
Document average age	4.72
Average citations per doc	8.738
References	393,489
Keywords plus (ID)	42,770
Author’s keywords (DE)	29,584
Authors	28,689
Authors of single-authored docs	2007
Single-authored docs	2894
Co-authors per doc	3.15
International co-authorships %	12.71

includes overall information about the articles indexed in Scopus, and it is seen that they are published in 4993 different sources (journals, conference proceedings, and books) in diverse subject areas from computer science, engineering, and social sciences through mathematics, decision science, medicine, business, management, and accounting. The different problems and contexts with which intelligent education is associated can be judged not only by the diversity of the subject matters of the sources but also by the large number of different keywords and references used by the authors. Articles prepared by a team of authors are about 14 times more than those with a single author. The percentage of participation by foreign authors (at least one of the authors must be from another country) is not high (12.71%), which indicates that collaborative studies are most often between authors from the same country.

A detailed contribution analysis of the top 10 countries shows that the greatest number of publications come from authors mainly from three countries: China, Germany, and the United States (Figure 2).

The other publications originate from countries, such as Brazil, India, Italy, Turkey, the United Kingdom, Spain, and Japan. China’s leading role is noticeable as the difference in scientific production between the first and tenth countries is about 30.61 times for 2024. Considering that the education system in each country is very specific, such quantitative differentiation may indicate a trend for further qualitative changes leading to significant differences in the practical implementation of IES in the future.

Regarding scientific content, the twenty most frequently used keywords by authors to describe the articles’ research focus for 2014–2024 are shown in Figure 3. Both technologies (e.g., artificial intelligence, machine and deep learning, natural language processing, data mining, learning analytics, intelligent tutoring systems, and virtual and augmented reality) and educational context (e.g., personalized learning, adaptive learning, higher education, and e-learning) are described.

It can be seen that intelligent education is associated with the processing and analysis of large amounts of data collected during the learning process, the application of techniques and algorithms from machine learning and artificial intelligence, and the development of relevant tools and software. Intelligent tutoring systems, as well as virtual and augmented reality applications, are also part of the implementation of intelligent educational environments. Other notable technologies are cloud computing and the Internet of Things.

Considering the term “intelligent education”, bibliometrix:biblioshiny application gives the trending topics for 2024—ChatGPT, Large Language Models (LLMs), and attention mechanism. ChatGPT is a chatbot developed on a large language model, GPT-4o, and is mainly used in question/answering systems and text summarization. Large language models are created to support tasks related to natural language processing. The attention mechanism is a technique in deep learning that is used to improve the accuracy of processing complex data. It can be said that a large part of the publications explores and discusses the application and impact of technologies such as ChatGPT in education, as well as various techniques for processing and analyzing complex data collected and used in the educational process.

The co-occurrence network shows two clusters with similar terms (Figure 4). The first cluster is formed around the keywords artificial intelligence, machine learning, and education (in blue), which has the biggest co-occurrence in comparison to other terms in the same cluster. The main keywords in the second cluster are intelligent tutoring systems and deep learning (in red).

A summary of the terms contained in the two clusters is presented in

Table 2. Groups of terms in the clusters.

Parameter	Cluster 1 (in Blue Color)	Cluster 2 (in Red Color)
Main keywords	Artificial intelligence, machine learning, education	Intelligent tutoring systems, deep learning
Technologies	Educational technology, virtual reality, augmented reality, information technology, computer vision, Internet of Things, cloud computing, robotics, chatbot, ChatGPT, gamification	Ontology, fuzzy logic, learning analytics, data mining, neural networks, natural language processing, knowledge tracing, reinforcement learning, knowledge graph, classification
Educational context	Higher education, engineering education, online education	Online learning, collaborative learning,
		personalized learning, adaptive learning, e-learning

, which are grouped into three groups: basic terms that occur most frequently, terms related to technologies, and terms that indicate the context of application of these technologies.

The overall, initial review of bibliometric information shows the general global picture in the field of intelligent education, which is associated with growing interest from the scientific community and an increasing number of research and publications. The most frequently used keywords by the authors outline technologies developed and used in different educational contexts, with trend terms for 2024 related to the research and use of large language models and chatbots.

4. Intelligent Educational Environments—Overview and Key Concepts

The authors consider the intelligent educational environment as an overarching concept that integrates aspects of both an intelligent educational system and an intelligent tutoring system. The nested view of the relationships among these systems is illustrated in Figure 5. As the earliest concept, intelligent tutoring systems can lie at the core of any intelligent educational system, providing means for personalized teaching and learning. The IES is intended to deliver educational services on a new qualitative level, thus facilitating the pedagogical process and managing the complex relationships between implemented technologies and IES users. In turn, it is part of an intelligent educational environment that connects it to the real world by sensing and controlling various environmental parameters. Each outer-level concept inherits the attributes and functionalities of the inner-level ones. The IEE is the broadest concept that reflects the interaction of the IES with the real world and supports the teaching–learning process.

The nested view of the correlations among the concepts of IEE, IES, and ITS shows their close connection and how technologies (can) enrich education, allowing for a more effective and entertaining experience. For example, personalized learning supported by ITSs through the use of AI and knowledge representation technologies can be implemented in any IES. Further, ICTs and intelligent technologies utilized by the IES give more opportunities for personalization. Adding functionalities to control and use environmental information to adapt the teaching process to the student’s needs and preferences turns an IES into an IEE.

Much research explores various issues related to intelligent educational environments. A more focused insight is given by analyzing the results of the query “intelligent AND education AND environment” in the database Scopus. It explores titles, abstracts, and authors’ keywords and returns 3472 documents for 2014–2024. The overview of the essential aspects discussed in the papers (Section 4.1, Section 4.2 and Section 4.3) is based on the trend topics of authors’ keywords in the query (Figure 6). These aspects are considered regarding the research questions and the current research focus.

As seen in Figure 6, artificial intelligence has been among the leading three keywords since 2021. This shows a stable interest in implementing AI in the educational framework. The presence of terms such as machine learning, deep learning, learning analytics, big data, cloud computing, and ChatGPT indicates the integration of intelligent technologies into contemporary education. This is also supported by the researchers’ keywords, such as online education, e-learning, and distance learning. The leading place of the term personalized learning for 2024 could be due to the synergy of these two groups of terms.

The returned query results are ordered by relevance according to the research topic. The most recent results are selected and further explored according to three main topics: IEE, IES, and ITS.

4.1. Intelligent Educational Environments

Intelligent educational environments are a radical advancement that drives educational innovations and transforms the organization and model from traditional to service-based [18]. The main idea of intelligent educational environments is to change the teaching–learning paradigm and provide a quality learning process and innovative educational resources to improve learners’ experience [19]. Thus, it is expected that improved quality of educational services and increased student engagement and outcomes will occur [20].

The relationship with the real world is implemented with the help of various IoT devices (e.g., sensors and actuators), which serve, on one side, to measure and control parameters of the microclimate (e.g., temperature, humidity, level of noise, light, etc.) [7,17] and, on the other, to assess the behavior of the involved users (e.g., teacher, tutors, and students) [5,6,21]. Technological support is essential for multiple reasons: data are generated and exchanged at a rapid pace, secure access control must be regulated, reliable data storage must be provided, and algorithms for monitoring and control must be developed and implemented.

Last but not least, users must be trained to work with new technologies and be willing to accept them. This represents a trend for the digitalization of services and interactions, which is part of the innovations in modern society. A digitally skilled workforce is part of the vision for Industry 4.0 and sustainable development in other economic sectors. There are also some associated risks, such as personal data leaks, the danger of unauthorized surveillance and access, and the increase in the transparency of human profile data [22].

Technologies integrated into the intelligent educational environment lay the groundwork for providing actionable feedback during teaching–learning interactions. Such feedback allows monitoring of the student’s engagement and enables teachers and tutors to personalize teaching to maximize students’ performance. From a pedagogical viewpoint, there are many approaches to implementing personalization, which are usually based on knowledge models.

Ontologies are used in many research projects in e-learning to enhance the personalization of learning. They provide a formalized structure for representing knowledge. In personalized learning, using ontology-based knowledge representation can help educational systems understand what learners already know and still need to learn. The system can recommend appropriate content, activities, or assessments by mapping learners’ current knowledge and skills to concepts defined in the ontology. The research in [23] proposes an ontology model, including both learner information and learning object properties for personalized content recommendation. The learner profile also includes dynamically changed data about the learner. Dynamic data are collected by tracking the behavior of learners and by using questionnaires and test results. An ontology for the description of learning objects based on the IEEE LOM standard is proposed in [24]. A competency ontology for learning environment personalization is proposed in [25]. This ontology is aligned with other ontologies used within the web of linked open data. The main idea for using mapped ontologies, modeling different elements of learning context for dynamic personalization in e-learning (including building and recommending adaptive learning paths and defining the prerequisite relationships between concepts), and variants of its realization are also discussed in many other recent publications, including [26,27]. Ontologies also can enable real-time feedback by identifying which concepts the student has or has not yet mastered and which concepts need more attention. This can lead to more effective and targeted interventions, such as recommending additional practice problems or providing hints specific to the student’s current learning needs.

4.2. Intelligent Educational Systems

According to Zhang and Cao, an intelligent educational system utilized for purposes of higher education should possess functions related to the identification of the presence and monitoring of students in the classroom as algorithms for face detection and face recognition are applied (based on convolutional neural networks), tracking the knowledge status through usage of deep learning (DL) memory augmented neural networks, and report generation for learning analysis [28]. These functions are based on DL algorithms, and in real time, the teachers could obtain such information in order to improve their teaching.

Zhang proposes a machine-learning method for forming a learning path to improve students’ experience and outcomes in personalized learning as part of an intelligent educational platform [29]. The presented approach performs better than other methods like collaborative filtering, neural networks, and graph clustering. The author concludes that this is a solution for the realization of effective personalized learning.

Zhu et al. talk about how to prepare and optimize an education system that integrates functions for personalized learning [30]. The factors considered for a personalized education system are utilized teaching strategies, massive resources, and a developed study and learning plan for students. The role of AI algorithms in optimizing resource planning (the particle swarm optimization algorithm is discussed) and improving system management (the suitability of the Decision Tree algorithm is commented on) is shown.

A framework for personalized recommendation of course resources that uses DL in the context of intelligent robot-driven education is presented by Li and Yang [31]. It is proved that the proposed approach that consists of the XLNet method, the Multi-Bi-LSTM algorithm, and the multi-headed attention technique is better in comparison to the so-popular algorithm of collaborative filtering. The performance of the recommendation mechanism is improved, especially in settings that generate and process big volumes of data.

To solve the problem of resource matching in intelligent education systems, Xiang et al. explore the possibility of a k-means algorithm to obtain similarity in users and resources [32]. They discuss the accuracy of the model created and its effectiveness and practicality in matching suitable resources considering users’ needs.

How to make an intelligent education platform effective at delivering educational resources and improving students’ satisfaction through AI technology is presented by Huang [33]. Utilization of AI has the potential to facilitate education management and support students’ training, leading to the realization of a high-quality educational process.

Chen talks about intelligent education, considering another aspect related to the improvement of data mining techniques [34]. A text analysis is conducted to outline the connection between knowledge and test questions. According to the students’ results obtained from the performed test, the learning characteristics of every student are described in order to group students with similar characteristics by applying the knowledge graph technique. The advantage of the proposed approach is its capability to reveal some issues in students’ learning and, in this way, contribute to achieving quality in teaching and the learning process.

Wei and Jin discuss an intelligent education management system based on IoT for optimizing course design [35]. Some of the classroom characteristics are presented visually in real-time, followed by the generation of information regarding learning status and personalized learning paths for every student. Some experimentations are performed, and the proposed system is verified to be suitable in the context of smart cities.

Another study is devoted to the realization of intelligent learning and functions related to its management through an Android-based voice assistant and wireless sensor network in higher education contexts [36]. Through voice instructions that are analyzed with machine learning algorithms, the teachers and students can accurately manage activities in the educational process and improve self-efficacy.

The benefits of virtual reality (VR) technology and education based on simulations are presented by Pense et al., who argue that intelligent transportation systems and VR can increase students’ motivation and improve educational quality [37]. The authors also mention some other VR advantages for students, like achieving better knowledge memorization and understanding, improvement of concentration, and learning performance.

Recently, ChatGPT and its application in educational settings have been widely discussed by educational society, showing the usefulness and drawbacks of this technology. Fütterer et al. explore the applicability of ChatGPT, arguing that it is a suitable tool for conducting conversations and giving the possibility to obtain experience with AI-based generative techniques and LLMs [38].

Adaptability and personalization are essential in intelligent education. During the last twenty years, intelligent technologies and knowledge representation formalisms have been successfully used to personalize education in the context of intelligent tutoring systems.

4.3. Intelligent Tutoring Systems

Intelligent tutoring systems (ITSs) were developed to support the direct application of AI to create personalized and adaptive learning. ITS lays the ground for intelligent education. They aim to facilitate teaching by modeling all the learning and tutoring processes (including tutoring knowledge) and emulating human tutor tasks, including knowledge presentation, problem statement interpretation, problem-solving process, and assessing the validity of the problem-solving process [39,40]. ITSs are computer-based educational systems that use various information about each student and knowledge models and utilize AI techniques to adapt the learning experience to the student’s needs [41]. Usually, ITSs provide personalized tutoring, learning, or feedback to learners, often in the form of a one-on-one tutor.

Much research discusses concepts related to ITS. Here is a summary of key issues and features of intelligent tutoring systems derived after a comprehensive analysis of scientific sources [39,41,42,43,44].

• Domain Model: It represents the learning content (usually in a machine-processable way) and provides the structure and knowledge needed to present lessons, problems, and feedback.
• Expert Model: The best possible model of expert knowledge in the domain from a scientific point of view that provides the system with correct answers, solutions, and reasoning strategies. It helps evaluate the students’ answers, i.e., the actual knowledge obtained, and identify gaps in their knowledge.
• Student Model: It represents the learner’s current knowledge, skills, psychological properties, preferences, and learning progress. It tracks individual performance and the evolution of preferences, learning styles, and behaviors.
• Tutoring Strategy/Pedagogical Module: It manages the overall teaching strategy and adaptation to the learner’s needs. It defines how the system presents and sequences the content, how feedback is delivered, and how it is used for personalization.
• Assessment and Evaluation Module: This module assesses students’ understanding of the material over time and evaluates their progress. It uses assessment tests, quizzes, or tasks to reveal whether and to what level the learner has achieved the learning objectives. Diagnostic tests, formative assessments, or summative assessments can also be included.
• User Interface: It interacts directly with users (students, teachers, or domain experts). It presents learning content, allows students to input responses, provides feedback, and ensures professional manipulation of knowledge models.
• Inference Engine: It interprets the data from the user input, the student model, and the domain model to make decisions about the next best actions (e.g., presenting problems, offering hints, adjusting difficulty). It uses logical reasoning or machine learning techniques.
• Feedback System: It provides immediate and constructive feedback based on the student’s actions and answers. Feedback can be corrective, encouraging, or suggestive (offering improvement strategies). It may also suggest hints and explanations or prompt the learner to think critically.

Thus, ITSs usually integrate several of the abovementioned components, enabling intelligent education. However, in practice, a variety of experimental ITSs were developed, implementing only partially some of the discussed components because of the complexity and difficulties in realizing one universal ITS.

Intelligent tutoring systems are relatively old for supporting personalized tutoring and learning, but their architecture is well-developed, very flexible, and adaptable to the requirements of modern intelligent education. Recently, significant research has been conducted on integrating ITSs with technologies based on modern large language models. Both the architecture of LLMs and ITS are used in Hypergraph-Based Intelligent Tutoring Systems and Conversation-based personalized tutoring systems. Hypergraph-Based Intelligent Tutoring System (HINTS) is an advanced type of ITS suitable for supporting problem-solving (e.g., in mathematics) [45]. This system applies Natural Language Processing techniques based on LLMs. A suitable LLM can also generate a source code solution (for example, in Python). Then, this source code can be transformed into ITS’ internal knowledge representation model. By intelligent automated control over the student’s reasoning and problem-solving process, this approach can help learners develop their reasoning, problem-solving, or math skills. The complexity of ITS is its main drawback. The development of a good ITS requires much time, labor, and qualified researchers. To simplify the development process, ref. [42] proposes a method for automatically converting mathematical problem statements from natural language to the internal system representation of an ITS. In a practical context, the researchers use large pre-trained language models to translate the problem into Python and then import this code into an ITS [46]. However, the integration of LLM-based systems into an ITS is still experimental because the accuracy is insufficient. Nevertheless, all advancements achieved in ITSs, including methods for ontology-based intelligent tutoring, apply to IESs.

5. Results

The bibliometric research and analysis of related studies revealed that the researchers have various views on IESs. Taking them into consideration, the authors create an overall conceptual model and further propose the IES to be part of an IEE that integrates IoT devices and environmental parameter control.

5.1. User Requirements for IEE

The first step towards creating the model of the intelligent educational environment is summarizing and formulating relevant requirements for the IES considering its application scenarios within the IEE [47]. To develop a conceptual view of an intelligent educational environment, the authors scrutinize many related papers, preferably focusing on recent systematic reviews and research concerning frameworks and key structure aspects of educational environments. Thus, based on critical analysis of the most relevant of them [7,8,9,10,13,15,16,18,19,21,24,28,29,35,39,41] and considering authors’ experiences in the field, the user requirements for the IEE are systematized and presented in

Table 3. User requirements for an intelligent educational environment.

Aspect of IEE	Users
	Students	Teachers	User-Independent
Intelligent Teaching–Learning Process	Discussion-enabling tools; Authoring task tools; Course selection tools.	Authoring resource tools; Tools for real-time group and individual interactions; Tools for feedback; Automation tools.	Collaboration and communication tools; Interaction tools; Tools for management of the teaching process; Assessment tools; Learning resources repository; External digital resource repository.
Technology (Educational and Environmental Technology)	Affordable, easy-to-use devices to access IES; Remote and mobile access.	Integration with other tools and resources; Tools supporting smart teaching: for creating a custom curriculum and personalized resources; Functionalities for innovative pedagogical technology-based approaches	Dependability; Availability of various educational technology tools and software; Standard-based interoperability of learning resources; Attendance registration. Networking; Connected devices (sensors, actuators): affordable and reliable; Capability to maintain optimal environmental parameters.
Data and Knowledge Management	Regular data gathering to enable appropriate and timely feedback; Data privacy and compliance with GDPR.	Tools for collecting, analyzing, reporting, and visualizing data and knowledge related to the individual student’s learning process; Statistical analysis tools; Tools for learning analytics, managing, and implementing smart teaching.	Tools for collection, storage, and analysis of the data gathered in IEE; Data mining, neural networks, and AI-based tools; Affordable data storage with high reliability; Compatible data structure, secure transmission, and storage.
System Management	Access rights; Personal space management.	Tools for extensive learning analytics, managing, and implementing smart teaching; Reporting services; Intelligent control and management of the whole IEE.	Functionalities for intelligent planning, organization, monitoring, evaluation, and control of the teaching–learning process within IEE; Compliance with institutional and security policies; Fault tolerance.

.

An IEE is a technology-rich educational environment that integrates various contemporary educational and smart technologies to support a full range of activities related to the teaching–learning process [3,7,19]. This environment involves specialized AI-based applications and software for managing all processes concerning data monitoring, collecting, and processing that support decision-making for achieving an enhanced learning experience. It utilizes innovative technologies to control both the intelligent teaching–learning process (including AI-driven applications, machine learning, learning analytics, ontologies, etc., to provide personalized and adaptive learning) [13,15,16,28,39] and smart learning workplaces (including environmental parameters and links to external devices and sensors to enable optimal classroom conditions and supply necessary external data) [6,7].

To maximize the benefits of an intelligent educational environment, teachers require it to provide tools for supporting smart teaching, such as authoring tools for creating personalized learning resources (educational content, tasks, tests, etc.), tools for composing individual learning paths, and custom curricula [12,29]. A communication interface is necessary to enable real-time interactions, collaboration, and discussions with peers or tutors within the IEE and to provide instant feedback [8,47]. Further, learning resources, courses, and curricula must comply with institutional requirements and related e-learning standards [18,24]. In order to do this, teachers have to receive from the IEE various learning process-related data and have analytical tools enabling deep insight into students’ learning performance [19]. These data should have an appropriate timeframe for gathering. The respective data storage should be affordable and capable of sustaining the necessary writing and reading speeds.

Various educational technology tools and software are necessary to support the IEE functionalities. To ensure sustainable work of the IEE, the technology in use has to be cost-effective, preferably based on open-source solutions and tools, easy to integrate and upgrade, and support the incorporation of new devices and technologies [8,21]. Different types of underlying networks (e.g., Ethernet, WiFi, and 3G/4G/5G) should be available and support connectivity among various devices (computers, smart screens, sensors, actuators, etc.) within the IES. The utilized technology has to be energy-efficient and highly reliable.

Based on continuous monitoring and learning data analytics embracing all students’ learning activities and achievements, an IEE enables smart knowledge delivery and assessment, smart collaboration, and constant feedback. Thus, the IEE represents an innovative concept that complies with the needs and requirements of students and teachers in the contemporary digital century by enabling personalized learning services.

5.2. Conceptual Model of an Intelligent Educational Environment

Considering the user requirements outlined in the previous section, a conceptual model of the IEE is proposed (Figure 7). It describes the entities that comprise an IEE and the relationships among them in the form of a UML class diagram. The entities are modeled by means of classes. One or more instances of each class may be created during the system’s lifetime. The class instances depend on each other and interact with each other as indicated by the different kinds of relationships represented by lines and arrows. The symbol “+” indicates public access to the methods while the symbol “#” indicates protected access to the attributes. Some classes that are central to providing the functionality of the IEE (e.g., SystemManagement, User, TeachingLearning, or Analytics) have multiple relationships with other classes. The SystemManagement class manages the green-colored classes Collaboration, TeachingLearning, Authoring, Assessment, and the gold-colored class Analytics (to improve the readability, some of these association names are not shown in Figure 7). Similarly, the User accesses functionalities from TeachingLearning, Authoring, and Assessment. Classes that model generic concepts, such as the class IoTDevice, also have multiple relationships—one for each subclass (e.g., LightingControl or AirConditioner). In contrast, subclasses and classes representing concepts that serve as building blocks of other concepts, such as the classes Resource or Exam, have fewer relationships.

A non-standard color coding is used to show the relationship of the individual class to the aspects of the IEE illustrated in Table 3. The classes that support the Intelligent Teaching–Learning Process are drawn in green color. The classes that are responsible for the integration of Technology in the form of various devices, tools, network implementations, etc., are depicted in blue. The classes related to Data and Knowledge Management are shown in red, while the classes that are part of System Management are drawn in purple. There is a single gold-colored class (Analytics), which provides functionality that is important for both the Intelligent Teaching–Learning Process and the Data and Knowledge Management.

At the heart of intelligent educational systems is the smart use of various ICT- and AI-based tools to enhance the teaching–learning process. They can adapt to individual students’ needs, provide personalized feedback, and track learning progress. Thus, one of the essential characteristics is personalization, i.e., the IES should adapt to each student’s learning pace and style by utilizing educational data to figure out where a student is struggling and adjust teaching approaches and the learning path and content accordingly. This functionality is realized through the TeachingLearning class together with all related subclasses (Figure 7). EducationalTechnology includes intelligent agents, educational data mining, learning analytics, ontologies, etc., which are successfully tested and applied in many ITSs. Knowledge-based technologies include models and tools that leverage knowledge representation and reasoning to solve problems related to personalized tutoring, recommendation of learning content, dynamic generation of personalized learning paths, etc. These technologies are designed to process information in ways that mimic human understanding and reasoning, enabling them to provide intelligent actions. Ontologies, intelligent agent-based systems, rule-based systems, case-based reasoning systems, or deep learning-based technologies can be used to ensure intelligent behavior of the intelligent educational environment, based on the proposed model.

The IES supports multiple learning modalities to cater to different students’ learning needs and preferences. TeachingLearning interacts with ResourceLibrary, which indexes and stores diverse content such as video, audio, interactive exercises, text, etc., and thus enables teachers to use various types of learning resources. Most of the functionalities of ITS are implemented by classes TeachingLearning, EducationalTechnology, Assessment, Analytics, and DigitalRepository. Also, the classes Authoring and Automation, together with EducatonalTechnology help teachers offer several personalization and adaptation modes. These modes are based on user profiles in combination with applied Analytics (e.g., learning analytics, machine learning, etc.). Further, feedback from Assessment also helps to refine personalization and adaptation approaches considering the results achieved.

The TeachingLearning and Assessment classes and users’ Profiles can store highly structured information, including ontological knowledge models. This information can form the basis of personalization strategies implemented by built-in multiagent systems. The EducationalTechnology class contains the attributes IntelligentTutors, Personalization, and KnowledgeBasedTech (e.g., ontologies, LLM, and machine learning) that can support the learning process of all students, including those with specific educational needs. In this way, they propose real-time personalization of the learning/tutoring process. Intelligent tutors usually work in multiagent environments, proposing built-in communication and collaboration mechanisms. They can use knowledge-based reasoning mechanisms to recommend the best learning resources to every student. Hybrid reasoning that combines rule-based, model-based, and case-based reasoning can be most useful in supporting intelligent tutoring. The class SmartEducationalDevice includes smart devices, such as interactive screens, web cameras, and various computer-based devices that are used in implementing the teaching–learning process.

Appliances that ensure classroom comfort, such as heating, ventilation, and air conditioning systems, are usually responsible for almost half the energy consumption in contemporary smart schools and universities. Thus, in integrated intelligent educational environments, one of the goals is to reduce operating costs. Modern appliances may be viewed as IoTDevice class instances in their own right, and these capabilities can be applied to control and optimize the microclimate in smart classrooms to achieve energy efficiency. Advanced smart technologies permit the continuous monitoring of appliances and can help achieve appropriate microclimate conditions in educational environments, as well as reliable control, fault detection, and hence—energy efficiency.

The conceptual model illustrated in Figure 7 incorporates the concept of gathering data from IoT devices organized in predominantly wireless sensor networks and controlling the learning environment via different kinds of actuator IoT devices. An IoT device (modeled by the class IoTDevice) may act as both a “Sensor” and an “Actuator”, which is why these device roles are represented as interfaces in the model. On a higher level, the data that are gathered by devices implementing the interface “Sensor” are divided into two broad categories—data related to the environment (temperature, humidity, air quality, luminosity, etc.) and data related to interaction with the participants of the learning process (e.g., presence of students, frequency of movements, gestural patterns, sound patterns, etc.). The data belonging to each category is managed by a separate class—Environment or Interaction, respectively. The devices implementing the interface “Actuator” also follow this categorization. Environmental control includes air conditioning systems, motorized windows, window shade control, smart lighting, etc. The actuators that are designed to enhance interaction include laser projection systems, multimedia systems with sound and video capabilities, smart physical models, laboratory equipment, etc. The reason for introducing these categories is that environmental management is relatively low-speed and usually does not need too much precision, while interactions tend to require a fast speed of both sensor measurements and actuator responses with a high degree of accuracy. In addition, environmental control is a traditional research area with many existing mature technologies, devices, and service providers. The topic of interaction is much more colorful and may incorporate a variety of different technologies often adapted from other sectors—e.g., industrial, automotive, or military technologies.

The system management of the intelligent educational system is tasked with the overall management of all sensor data and IoT device interactions. The Environment and Interaction classes are tasked with keeping track of the past, present, and desired future states of the environment and the interactions, respectively. The communication with the devices takes place via specialized communication hardware (class IoTGateway), which provides a communication bridge between the IoT device networks and the intelligent educational system. An IoT device data storage capability (class IoTDataStorage) is also envisioned to support analytics and decision-making.

The intelligent educational environment is technology-packed and, like any technical environment, must consider dependability and information security aspects, as well. Dependability is a concept usually associated with high-reliability and safety-critical systems [48,49,50], where the consequences of system failure could be catastrophic. Therefore, they are designed and built to resist external and internal faults. The high-reliability requirement brings additional costs, and a trade-off has to be made between dependability and expenses. The educational environment does not impose heavy fault-tolerance requirements on the applied technical solutions. Some dependability concepts and approaches, however, are applicable in modern intelligent educational environments. Continuous and flawless operation can be achieved by replicating the most valuable system assets, applying error detection and recovery of the data, and using fault-tolerance techniques for real-time operations. For example, the IoT gateways and device data storages may be replicated to enhance the reliability and availability of their functionality (represented by the attribute Replicas in classes IoTGateway, IoTDeviceDataStorage, and DigitalRepository). Error detection techniques may be applied to the data acquired from the sensors.

Some vital IES subsystems, such as data repositories, system management, and IoT device control, must be fault-tolerant. Their reliability can be achieved by hardware or software redundancy, design diversity, error detection, etc. Assets and functions associated with real-time (like online examination) should envision techniques such as encryption, signature monitoring, duplication, etc. Fault-tolerance approaches, in combination with information security measures, can assure information protection and continuous functioning of the IEE.

Information security is a significant aspect that should be taken into account in the IEE. Information security techniques are essential in data creation, storage, management, and transfer [51]. Techniques for data protection, authorized access to information, availability of the needed learning resources, protection of sensitive data, prevention of tampering with data, and the theft of sensitive information are some of the security measures needed. They include methods such as encryption, anonymization, data erasure, and data backup. For example, control of access rights and conformity to the GDPR act [52] are included in the class SignIn. Security techniques should be incorporated during the design of the IEE and comply with security policy [53].

Dependability and information security need special attention in the context of an IEE. They permeate almost every element of the proposed UML model without being a single class. As suggested in [54], their integration into the IEE should be envisioned in the IES’s design or even its conceptualization.

5.3. Practical Applications of Intelligent Educational Systems

The given examples illustrate how the proposed conceptual model can be practically implemented. They demonstrate some aspects of the IEE and explain the relationships among particular classes of the UML model when providing specific functionalities such as student assessment and personalized teaching and learning.

5.3.1. Personalized Teaching and Learning

For the automatic conduct of personalized learning and tutoring, the IES stores knowledge about learners in profiles (class Profiles) and manages this knowledge. Several mechanisms for storing structured knowledge and methods of its usage in organizing the tutoring process (reasoning mechanisms) are used. Various knowledge modeling methods are applied to develop students’ models, domain models, tutoring, and expert models in IES, including ontology-based, knowledge graph-based methods [55], mathematical or simulation models, and data-driven methods. Knowledge-based models have built-in or use reasoning mechanisms, including model-based reasoning [40], case-based reasoning [56], and rule-based reasoning [57]. A combination of these approaches is usually applied.

Ontology-based knowledge models used by the class EducationalTechnology are the primary way to store knowledge in a machine-processable way and, in this way, support automated reasoning. Ontologies are used as knowledge sources, supporting the following tasks in IESs: (i) natural language processing (for text analysis or students’ natural language input) [57]; providing (ii) adaptive learning paths (based on ongoing assessments); (iii) personalized and adaptive teaching–learning; (iv) domain knowledge representation allowing explicit revealing of the relationships between concepts; (v) improved semantic understanding by utilizing the structured relationships in ontologies; and (vi) content generation, annotation, recommendation, and reuse (class Authoring). The described tasks are implemented by TeachingLearning and EducationalTechnology classes with the help of ResourceLibrary and DigitalRepository classes (Figure 7).

Personalization in an intelligent educational system allows adaptation to each student’s learning knowledge, skill level, needs, and style. Personalization of the teaching–learning process includes mapping the student’s profiles (learning styles, preferences, skills, etc.) with learning resources. It should be dynamic in the context of changes in some student characteristics, such as knowledge level and learning needs. Hence, an appropriate tutoring model for storing these characteristics and performance in the learning process is essential. The system collects data on student participation, time spent on tasks, specific tasks or overall performance, etc. The integrated IoT devices such as motion sensors, cameras, and microphones regularly provide data for the presence and behavior of students in the smart classrooms, i.e., for their involvement in the educational process. In the UML model (Figure 7), these devices are represented by the base class IoTDevice and its subclasses, which connect to the system using instances of IoTGateway. These data are used to timely inform both students and tutors to help them identify trends, predict outcomes, or suggest interventions. The IES utilizes data gathered within learning interactions (class Interaction), processes them using the class Analytics, which supply learning analytics algorithms and other intelligent technologies, and provides them to the TeachingLearning and EducationalTechnology classes. The latter handles intelligent tutoring, adaptation, and personalization. This functionality enables the IES to obtain insights into students’ strengths, weaknesses, and learning patterns. Thus, the students’ performance is assessed to determine where they are struggling, and the educational approaches, learning content, or other parameters of the teaching process are adjusted accordingly. Based on the students’ learning interactions and outcomes, the system can change the applied teaching approach or activities and recommend specific educational resources from ResourceLibrary and DigitalRepository utilizing ontology representation of knowledge. For example, if students cope with some problems/tasks quickly, they might be presented with more challenging ones; otherwise, the IES could offer simpler tasks or supplementary explanations.

Furthermore, thanks to the Assessment and EducationalTechnology classes, users receive immediate feedback on the learning process. This helps students understand their mistakes and correct their misconceptions. In addition, teachers are offered support for quick adaptation of learning approaches, activities, materials, etc. Overall, student performance data are reflected in their profiles to update their knowledge levels and other attributes.

5.3.2. Assessment Process

For the assessment of learners’ achievements, the Assessment class plays a key role. Assessing is a mandatory and extremely important component of the formal educational process. It shows how successfully or unsuccessfully they have achieved the learning objectives, improved existing knowledge and skills, or acquired new knowledge and skills. The assessment results can be used not only to improve the learning process but also to enhance teaching by providing valuable feedback to students and teachers. Diagnostic assessment is used to evaluate students’ knowledge before beginning the course or any learning unit. Thus, it gives prompt feedback to teachers and supports choosing a better approach to provide and distribute new knowledge. Formative assessment is applied to assess the planned learning tasks during the course, stimulating students to be active participants all the time, which is to their benefit, and helps in preparing for the final exam. A summative assessment is conducted to obtain a midterm or final exam grade. The Assessment class allows a wide variety of assessment tasks and different assessment types to be realized, providing flexibility for the teachers considering the assessment strategy that they have to implement.

Self-assessment is an important learning strategy in which the learner controls self-learning, learning progress, and output results. It is a way for the students to understand their knowledge gaps and improve some competencies or skills in informal educational settings. Self-assessment can also be used as a supplementary tool to support formal education. The Assessment class can also include self-assessment tasks, and statistics of the assessment results can be included in the Analytics class. The analysis of the student’s self-assessment can be further compared with the teacher’s opinion, and the result can be used to detect gaps in the assessed knowledge and skills. Comparing the marks awarded by the students themselves when performing specific assessment tasks with the teacher’s grades leads to the identification of inappropriate strategies for conducting training and teaching. This will be a reason for additional analysis and improvement of the educational scenarios used.

Assessment has been significantly facilitated for the participants in the educational process because several machine learning and artificial intelligence techniques are applied in IES and included in the class EducationalTechnology. A quick and objective assessment can be obtained for multiple types of assessment tasks, starting from a standard exam test and moving on to the automated evaluation of free text in the form of an essay or project and reaching the automated assessment of graphic objects (in the form of graphs, diagrams, images, etc.). This contributes to the effectiveness of teachers, especially when personalized learning needs to be implemented, and at the same time, they must assess a large number of learners.

The learners’ actions in the assessment process are monitored as the environment collects and processes data, enabling an analysis of the achieved learning outcomes and progress. For this reason, in the conceptual model of IEE, the Assessment class uses the functionalities of the Analytics class to gain deeper insight into students’ knowledge and achievements. This analysis is an opportunity to present and summarize the current status of the assessment tasks (completed or pending) and their timeframe. It contributes to improving the learner’s efficiency, taking into account the planned activities. The classes Assessment, Analytics, TeachingLearning, and the subclass Student implement the program logic for choosing the appropriate learning activities, particular learning path, or pedagogical approach necessary to improve students’ learning outcomes. In the context of self-assessment, the intelligent environment provides opportunities for the learner to understand their current knowledge better, whether the completed tasks were appropriate for obtaining the relevant learning outcome, and how to plan future activities.

The Analytics class contributes to processing and analyzing data gathered through the assessment process. It is always helpful for learners to understand their progress and for teachers to improve assessment tasks and teaching strategies as a whole. Intelligent analytics is characterized by the capability to give prognoses predicting future events based on historical assessment data or data collected in real time. Such predicted events could be the next mark, the final outcome from the assessment activity, the drop rate of learners, etc. Supervised machine learning algorithms like Decision Tree, Random Forest, k-NN, and artificial neural networks are mainly applied and can make predictions with high accuracy. Large language models and ChatGPT have recently gained popularity and provoked extensive discussion about their usefulness and applicability in the educational context. The ethical aspects of these new technological solutions are also under debate, pointing out some negative influences on assessment and assessed learners.

6. Discussion

6.1. Answers to Research Questions

The presented conceptual model of an IEE aims to systematize the relationships among the ICTs applicable in education and the learning-teaching process and provide insights on how they could benefit each other. The authors’ vision is by no means exhaustive or complete. The UML conceptual model encompasses the main elements of an IEE and their interactions in the learning environment.

Intelligent educational systems, in some sense, are successors of well-known intelligent tutoring systems that also use artificial intelligence technologies for personalization. IES is a broader concept that emerged somewhat later and includes not only personalized learning and tutoring but also other elements like learning management systems, various attractive learning resource types, specific assessment tools, educational data analytics tools, and learning analytics. The IES also includes a variety of tools, systems, and applications that enhance not only the teaching and learning but also the administrative, assessment, and management aspects of education. ITS can be considered a subcomponent of a broader IES. Also, ITS or some of its components can be integrated into IES to provide a personalized and adaptive learning experience.

To achieve the research goal, the authors formulate four research questions and attempt to provide answers to them. The current paper outlines recent research trends in intelligent education and the impact of modern technologies on the contemporary teaching–learning process. The Scopus scientific database indexes a wide range of research fields, including pedagogical topics, engineering aspects, and computer-related subjects relevant to the research. Therefore, this database is appropriate for conducting an interdisciplinary study like the one presented in this paper.

RQ1: What are recent topics and research trends regarding intelligent education?

The bibliometric analysis reveals a huge increase in the research activity in intelligent education worldwide during the last decade (Figure 1). Most explored are the top 20 terms, based on the authors’ keywords (Figure 3). After qualitative analysis, they can be distributed into the following four groups: (1) intelligent data processing and knowledge retrieval technologies (artificial intelligence, machine and deep learning, natural language processing, data mining, learning analytics, and big data), (2) supportive technologies (cloud computing, virtual and augmented reality, and the Internet of Things), (3) educational context (higher education, personalized and adaptive learning, and e-learning), and (4) intelligent tutoring systems (as a generative term).

RQ2: What are the main user requirements for intelligent educational environments?

The authors review and summarize relevant publications on the topic and derive the most significant user requirements for IEE (Table 3). They are categorized as user-independent and specific (considering students’ and teachers’ views). On this basis, four groups of user requirements are outlined regarding the main aspects of the IEE—Intelligent Teaching–Learning Process, Technology (including educational and environmental), Data Management, and System Management.

RQ3: What essential elements and functionalities should be integrated into the conceptual model of an intelligent educational environment?

A generalized view of the essential elements is presented in the UML diagram (Figure 7), whose classes are related to the user requirement groups for the IEE defined in Table 3. The Intelligent Teaching–Learning Process includes all the teaching–learning activities (authoring, tutoring, learning, assessment, discussing, feedback, etc.). In the UML model, they are supported by the green-colored classes TeachingLearning, User (and its sub-classes), Profile, Authoring, Assessment, Collaboration, EducationalTechnology, Automation, and the gold-colored class Analytics.

Technology embraces both educational and environmental devices. From an educational perspective, here are all ICT devices and tools, such as networked computers and tablets, smart boards and screens, projection systems, video and audio devices, 3D printers, etc. To consider aspects of an educational environment and students’ behavior, smart devices that ensure appropriate and comfortable conditions for providing an intelligent educational process are included. Furthermore, these devices can support innovative teaching tasks by supplying environmental and behavioral data. The UML model represents these capabilities by introducing the blue-colored classes SmartEducationalDevice (and its subclasses), IoTDevice (and its subclasses), IoTGateway, Environment, and Interaction as well as the blue-colored interfaces Sensor and Actuator.

Data and Knowledge Management is responsible for the proper and secure creation, storage, processing, and transfer of data throughout the environment. The UML classes associated with data and knowledge management in the model are depicted in red and gold colors: DigitalRepository, ResourceLibrary, ExaminationLibrary, Resource, Exam, Content, AudioContent, VideoContent, TextContent, Analytics, and last but not least, IoTDeviceDataStorage.

System Management organizes the proper work of the whole IES and its interactions with the IoT devices through the gateway. It is responsible for the smooth operation of all functionalities of the intelligent educational environment, especially for organizing the intelligent educational process. The UML classes associated with this functionality are shown in purple color: SystemManagement and SignIn.

RQ4: What are the practical applications of the proposed conceptual model of an intelligent educational environment?

Two possible application examples that concern personalized teaching and assessment based on intelligent technologies within IES are discussed in Section 4. These examples illustrate how the proposed conceptual model can be practically applied. They demonstrate some aspects of the IEE and explain the relationships among particular classes in the UML model to provide these specific functionalities.

This paper represents the initial steps in the development process of an intelligent educational environment. It starts with bibliometric research and analysis limited to the Scopus database, the period of 2014–2024, and the particular queries considered relevant to the manuscript’s research goal. The obtained results serve as guidelines for the formulation of user requirements and basic elements of the conceptual model of the IEE. At this point, the proposed model has not been empirically and entirely tested. Only some aspects are studied and implemented. An IEE is complex, with many interrelated elements that interact in various ways. These features make the IEE validation a difficult task, which is part of future study.

Introducing modern technologies in today’s schools and universities imposes challenges of different kinds—social, cultural, ethical, economic, etc. There are countries or regions in the same country that lack internet connectivity. Building a technology-intense classroom requires significant, unaffordable costs for some countries. At the same time, the pace of ICT incorporation into the educational world is increasing and seems inevitable. That disproportion is disadvantageous for the countries that invest less (or no) resources into modern education technologies. However, they could benefit from the experience of other states, which are already developing the integration of innovative technologies in education. On the other side of the spectrum is the excess of modern devices that could lead to, for example, health problems and frustration, laziness in searching and acquiring knowledge, and distancing from the real world. All these significant challenges are worth investigating but are out of the scope of the current research. The authors strongly support the reasonable balance between technology utilization and traditional pedagogical methods.

6.2. Future Study

Part of the future study that the authors envision will include the creation of a prototype IES according to the conceptual model discussed in this paper. A strength of this implementation will be the integration of various IoT devices that will support the educational process along with more traditional computer-based devices. The research conducted so far shows that IoT devices developed for other industries may be good candidates for adaptation and integration into the new system if their traditional use cases include environmental monitoring or imparting new information on human actors.

As another aspect of the future study, the authors consider developing an e-learning personalization ontology that models all concepts related to e-learning personalization and their relationships, thus supporting interoperability between all components needed for content-based personalization. It is the best way to semantically model relationships between all the elements of intelligent educational systems. This ontology should organize concepts related to several classes of the UML model, including User, TeachingLearning, EducationalTechnology, DigitalRepository, ResourceLibrary, etc. Such an ontology would be very useful for proposing a general model of concepts, relationships, and rules related to personalized e-learning. This ontology would support creating a shared understanding that assists in delivering personalized content, adaptive learning paths, and customized assessments.

7. Conclusions

The contributions of this article may be summarized as follows: (i) bibliometric research of recent-decade scientific studies on intelligent education and intelligent educational environments; (ii) analysis of the user requirements for an IEE; (iii) development of a UML model of IEE; and (iv) practical applications based on the developed UML model.

Intelligent education has been a fast-developing scientific area in the last decade, and the trends show a growing interest in the following three fields: large language models, chatbots, and the development and application of a great range of technologies in education. This intense presence of ICT and intelligent technologies in today’s classrooms changes the educational paradigm and necessitates a structured view on their inclusion into the teaching–learning process for the benefit of all participants. Modern educational systems tend to become educational environments that offer personalized experiences to the learners and teachers and interaction with the real world.

Considering the key issues outlined by the bibliometric analysis, the authors systemize the users’ requirements for an intelligent educational environment and take them into account during the development of a conceptual model. The proposed UML model is a general description of an intelligent educational environment and does not include all its possible functionalities and classes. It gives the overall picture of the relationships among the essential elements of the IEE and provides insights into its future development. Some applications are described to illustrate how the model may be employed in a practical setting using the proposed concepts. The relationships among the IEE’s elements are complex, and any particular application of the model to build an IEE must consider all specific user requirements and add relevant details.