Skip to Content
MDPIMDPI

Advanced

ElectronicsElectronics
PDF
  • Review
  • Open Access

26 April 2026

49 Pages

Combinations of Generative Artificial Intelligence and Robotics in K-12 and Higher Education: A Review

and
Department of Education Sciences in Early Childhood, School of Education Sciences, Democritus University of Thrace, Nea Chili, 68131 Alexandroupolis, Greece
*
Author to whom correspondence should be addressed.
Electronics2026, 15(9), 1835;https://doi.org/10.3390/electronics15091835
This article belongs to the Special Issue Pushing Boundaries: Innovations in Robotics, Artificial Intelligence, and Extended Reality
Version Notes
Order Reprints

Abstract

Artificial Intelligence (AI) and robotics constitute two major technological fields frequently integrated into education. Both of them provide advantages to educational settings, stemming from approaches integrating them at all educational levels. The emergence of generative Artificial Intelligence and the growing popularity of related tools have accelerated the integration of AI into education. An aspect of interest is to explore the combination of AI with robotics in education, aiming to benefit from the advantages of both technological schemes. This paper reviews work regarding the combination of generative Artificial Intelligence and robotics in K-12 and higher education. Scopus was used to search for relevant work. Fifty-four relevant papers were retrieved and analyzed after an exhaustive search. Trends in this combination are highlighted, taking into consideration learning, teaching, robot functionality and capabilities of generative AI tools, teaching subjects, sample size, and educational levels. Five main types of generative AI and robotics combinations are discerned. The overall combination benefits and challenges are analyzed. To the best of the authors’ knowledge, there is no other review discussing this subject in this specific context.

1. Introduction

For decades, efforts have been made to integrate technology into education, with the primary goal of enhancing learning outcomes. Various approaches have been implemented across all educational levels. Effective integration requires careful planning that considers the needs of students and educators, curriculum requirements, and available infrastructure [1].
Technology takes many forms and offers diverse features, which should be taken into account when designing technology-enhanced educational strategies. Among the most prominent fields commonly incorporated into education are Artificial Intelligence (AI) and robotics.
AI is integrated at all levels of education from early childhood to higher education ([2,3,4,5,6]). Specifically, the integration of AI in higher education spans several decades. One may discern different directions in AI integration in education. More specifically, AI in education may be a teaching subject, learning and teaching medium, and a tool for learners, educators, and institutions.
In higher education, AI has been a teaching subject in computing and engineering departments for many years. It is a teaching subject in other departments as well, and interest is growing due to the increased popularity of AI applications. As a teaching subject, AI may be taught independently or in combination with other subjects (e.g., robotics). Both directions may be pursued at undergraduate and postgraduate levels. Recently, there has been a growing interest in integrating AI as a teaching subject in K-12 education as well, usually in the overall context of technology.
As far as learning is concerned, AI offers advantages as it provides personalized learning experiences by adapting content and instructional methods to individual needs. In the context of educational institutions, efforts are made to develop AI-based learning systems accessible through the web [7,8]. These systems are developed and supported by the institutions. The recent availability of popular AI-based systems on the web and the incorporation of AI methods in many systems provide further opportunities to learners at all educational levels.
AI provides opportunities in teaching and research. AI may assist in creating and refining learning content and tools. AI may also assist in the design and implementation of learning activities [6]. Educators can also process and analyze student data with the assistance of AI. This assists them in obtaining information about students’ overall progress, the difficulties they face, and potential gaps in their knowledge [9]. As far as research is concerned, AI may have multiple roles. On the one hand, AI may be the subject of research by itself or in combination with other subjects (e.g., robotics). On the other hand, AI may serve as a tool assisting research in all subjects.
The emergence of generative AI has expanded the benefits that AI can offer to students and educators, reshaping how they perceive its role in education [10,11]. The main advantages offered by generative AI include the ability to generate and refine content through natural language interaction. Interaction may be text-based, but other forms may be used as well. Generative AI has the potential to transform the aforementioned roles of AI in education. An interesting aspect is that these roles may be performed in combination with robotics.
Robotics may have multiple roles in education due to the variety in robot forms and functionality. Robotics in educational contexts may be a teaching subject, a learning tool, and a learning and teaching medium. Furthermore, it may provide services to learners, educators, and institutions.
Robotics is a teaching subject in several higher education fields and provides the opportunity for cross-curricular activities involving computing, physics, mathematics, and engineering [12]. It is taught at both undergraduate and postgraduate levels. It is also a teaching subject in K-12 education and often integrated with other subjects in the context of learning activities [13].
Robots may constitute useful learning tools. Students are able to take part in hands-on experiences by designing, building, and programming robots [14]. Students may also learn by experimenting with robots and observing their reactions. The implementation of robotics projects assists in promoting creative thinking, problem-solving skills, and collaboration [15]. Abstract concepts are put into practice in an engaging way and become more understandable [16].
Robots may assume further roles in education besides being learning tools or teaching subjects. This may be achieved with social robots that have advanced capabilities. Social robots are designed to interact with students and educators in engaging and meaningful ways. The capabilities of social robots are discussed in [17]. Social robots have perception abilities, including vision, audio, touch, and proximity. This enables them to recognize the environment, objects, and humans around them and respond to what is happening. They may comprehend natural language and interact accordingly with humans, perhaps with limitations concerning their ability to participate in open-ended conversations and to respond to open-ended questions effectively. They may recognize non-verbal human behaviors and respond accordingly. In addition, they may recognize verbal and non-verbal emotions of humans in their interactions with them. Furthermore, they may also express emotions themselves while interacting with humans. They also have the ability to connect with other systems. Due to the aforementioned capabilities, social robots can assume a variety of roles in educational contexts, including teacher, teaching assistant, companion, or friend [18].
Various approaches have explored the combination of AI and robotics in education. The goal of these approaches is to combine the features of both AI and robotics in educational settings. For instance, AI modules may be encompassed in (or collaborate closely with) robots, improving their functionality. However, other combination schemes may be explored as well. More specifically, AI tools may support learning, teaching, and research in robotics. Furthermore, AI-based systems may teach robotics.
The recent advances in generative AI have affected the research carried out regarding the combination of AI and robotics. More specifically, a recent research direction is to explore the combination of generative AI and robotics. Based on the aforementioned, one way of combining them is to use generative AI to improve the functionality of robots. In the case of social robots, it enables them to enhance their interaction in natural language and comprehend natural language in more contexts, surpassing previous limitations [19]. This improves their aforementioned roles in education as tutors, teaching assistants, companions, and friends. In the case of robots that are not social, their functionality may be improved with generative AI, enabling them to respond to human instructions and constitute enhanced learning tools. Similar to the aforementioned, generative AI may support learning, teaching, and research in robotics and may also teach robotics. It would be interesting to study all these aspects in the context of K-12 and higher education.
This paper reviews approaches combining generative AI and robotics in K-12 and higher education. Scopus was used to retrieve relevant studies. The review highlights combinations focusing on various aspects involved in education. More specifically, five main aspects of the combinations are discerned. These aspects concern robot functionality, AI support for learning in robotics, AI literacy, AI support of teaching and research in robotics, and AI-based teaching of robotics. A subset of these aspects is usually discussed by other researchers working in the field of generative AI and robotics in K-12 and higher education. However, it is necessary to consider all of these aspects in the discussion of the corresponding combinations.
This paper is organized as follows. Section 2 provides background knowledge. Section 3 presents the methodology, pointing out the research questions and how the search for relevant work was conducted. Section 4 presents the main trends in the surveyed work. Section 5 provides the discussion. Finally, Section 6 concludes.

2. Background

This section provides background knowledge for the following sections of the review. It addresses the following: (i) robot categorization and robotics in education, (ii) generative AI applications and challenges, and (iii) factors affecting the integration of technology in education.

2.1. Robot Categorization and Robotics in Education

Different categorizations for robots may be defined as they can be distinguished into categories using various criteria. For instance, a criterion for categorizing robots may be the environment they operate [12]. In this case, robots may be distinguished as fixed and mobile robots. Robotic arms usually associated with industrial tasks are examples of fixed robots. Mobile robots may move around their environment, and they may further be discerned as airborne, aquatic, and terrestrial ones. Typical types of terrestrial mobile robots include wheeled, tracked, and legged robots [12].
Another approach would be to categorize robots according to their application field and the corresponding tasks they perform. In this case, one may discern industrial and non-industrial robots [12]. Typical cases of industrial robots are robots for manufacturing and logistics. Typical cases of non-industrial robots include medical, home, educational, defense, space, and agricultural ones.
Some of the reviewed studies address robotic surgery; therefore, certain aspects of medical robots will be outlined. Medical robots have been used in surgery for several years, enabling doctors to operate remotely and/or with greater precision. In addition to surgical applications, medical robots can perform a range of other tasks, including rehabilitation, assistive wearable functions (e.g., prosthetic body parts), cardiac assistance, gastroscopy (e.g., capsule robots), magnetic actuation, and bronchoscopy (e.g., continuum robots) [20].
Various types of educational robots may be discerned. One way to discern them is according to the role an educational robot has. On the one hand, an educational robot may be a learning tool; that is, the robot has a passive role [13]. In this case, students mainly implement tasks such as designing, building, and programming robots. On the other hand, an educational robot may have a more active role; that is, it may assume the role of a tutor, teaching assistant, companion, or friend [13]. These last roles are assumed by social robots whose advanced functionality enables them to interact socially with humans.
Various robots are used as learning tools depending on the educational level [21,22,23]. In early childhood education, robotic devices in the form of floor robots are frequently used, enabling students to learn sequential logic and directional concepts. In primary education, robotic activities usually address simple block-based programming, computational thinking, and simple sensors and motors [13]. In secondary education, more advanced building, electronics, and programming are involved. In higher education, a variety of robots may be employed based on the subject of the studies. Tasks carried out in higher education may vary from introductory to advanced robotics, including real-world applications, AI methods in robotics, and autonomous systems. Projects focusing on collaboration, problem-based learning, programming, and hands-on experience are frequently implemented in the context of robotics courses [14,15]. Robot kits are frequently used in educational settings to provide hands-on experience as students design and build their own robots using the available components. Programming tasks in robotics involve screen-free activities and visual programming in the younger ages and text-based programming (e.g., Python and C++) for older students [13].
In several cases, robotics simulators are used in education. This is done at all educational levels but mostly in higher education [24]. Simulations are useful to reduce costs in cases of physical robots, which are expensive to acquire or difficult/impossible to use in educational settings. They also enable testing, experimentation, and practice using diverse scenarios without the risk of damaging robots and with safety for humans.
Challenges need to be addressed in the integration of robotics in education [13]. These include the costs associated with the required hardware and software and the need for curricula and infrastructure to adhere to the continuous advancements in the corresponding technologies. A further challenge stems from the interdisciplinary nature of the field, which creates the need to coordinate learning outcomes and content in multiple courses.

2.2. Generative AI Applications and Challenges

Generative AI focuses on AI approaches that generate new and diverse content in response to corresponding inputs. These approaches are typically trained on large datasets and, in many cases, also learn from web-based data [25]. Generative AI systems are commonly accessed via the web by connecting to remote servers. They may also be deployed locally when models are open, the available hardware satisfies model requirements, and sufficient technical expertise personnel are available. Generative AI systems available from remote providers may have a free version and versions requiring subscription.
The available generative AI models have been trained with datasets chosen by their developers. It may be possible to train available generative AI models with specialized datasets chosen by the researchers in educational institutions. In this way, the generative AI models are adapted to the needs of a specific domain.
The inputs and outputs of generative AI systems can take various forms. Consequently, generative AI approaches can be categorized based on their output type (e.g., text, images, 3D models, and videos) [25]. This section primarily focuses on text generation, as it is the dominant form of content in applications combining generative AI and robotics. Other types of generative AI are also briefly outlined, particularly those referenced in the reviewed studies.
Text generation systems produce natural language using language models. Large language models (LLMs) are the most prominent examples. Widely used LLMs include OpenAI GPT (Generative Pre-trained Transformer), Google Gemini, Claude, DeepSeek, and LLaMA. These models underpin conversational AI applications such as chatbots, which provide user-friendly web-based interfaces. Among these, ChatGPT is the most widely used, followed by alternatives such as Gemini, Microsoft Copilot, Claude, and DeepSeek [26]. In addition to LLMs, smaller language models also exist; these require fewer computational resources and are more suitable for local deployment. This direction has been explored in some of the reviewed studies.
While LLMs primarily process and generate text, many also support multimodal inputs such as images and audio. Multimodal models enable interaction across different data types, increasing flexibility. Vision-language models, for example, combine visual inputs (images and/or videos) with text and are commonly used for tasks such as image captioning, visual question answering, and object recognition [25].
Image generation approaches create new images or modify existing ones. Popular tools include DALL·E, Adobe Firefly, Midjourney, and Stable Diffusion, the latter being open-source [25]. All of these tools have been utilized in the reviewed studies.
Other forms of generative AI include, among others: (a) video generation from text, images, or videos; (b) text-to-speech systems; (c) code generation; (d) music creation; and (e) 3D model generation from text, images, sketches, or drawings [25]. Relevant tools from the last four types used in the surveyed studies are presented in Section 4. As mentioned in Section 4, researchers may implement their own customized generative AI tools for specialized tasks addressed in their studies.
Applications developed in educational institutions may integrate generative AI services from remote providers (e.g., LLM functionality) using APIs (Application Programming Interfaces). APIs enable external applications to submit prompts and receive the results. API metering costs may be charged based on API usage. There is also an API token limit referring to the amount of data that may be sent to or received from APIs [27]. APIs may also be used in case of generative AI services running locally.
Generative AI applications introduce several challenges, as discussed in [28], across four main perspectives: (i) ethics, (ii) technology, (iii) regulation and policy, and (iv) economy. The most critical challenges related to the first three perspectives are outlined below, as they are particularly relevant to education. Further details are available in [28].
Six key ethical challenges can be identified [28]. First, generative AI systems may produce harmful or inappropriate content. This concept is context-dependent and may include issues such as violence, offensive language, or culturally inappropriate material. Second, generated content may reflect biases related to gender, race, nationality, language, or culture. Both challenges are closely linked to the quality and diversity of training data and the web-based sources.
Privacy and security also present significant concerns [28]. Data from user interactions may be incorporated into training datasets or exposed to unauthorized access. Another challenge is the potential misuse of generated content by students, such as submitting AI-generated work as their own or using it in online assessments. Additionally, there is a risk of over-reliance on generative AI, where users may depend on generated outputs without critical evaluation. A further issue is the digital divide, which refers to inequalities in access to and use of digital technologies [29]. Limited access to generative AI may result from inadequate hardware, poor Internet connectivity, financial constraints, or insufficient user knowledge.
Technological challenges are also significant [28]. First, obtaining high-quality and sufficiently large training datasets remains difficult, which can negatively affect output quality. Second, generative AI systems may produce hallucinations, including factual inaccuracies or fabricated information. Third, the decision-making processes of these models are often opaque, making it difficult to explain how outputs are generated. Explainability is particularly important in domains such as education, as it enhances transparency and trust [3]. Fourth, verifying the authenticity of generated content remains challenging. Finally, output quality is highly dependent on effective prompt design. Prompt engineering (i.e., the process of crafting and refining inputs) requires skill and experience.
Regulatory and policy challenges are also evolving in response to rapid advancements in generative AI [28]. This is especially relevant in education, where regulations and policies must adapt to address the implications of AI use in teaching, learning, and assessment.

2.3. Factors Affecting the Integration of Technology in Education

Several researchers have studied the factors that affect the integration of technology in education. The identification of the corresponding factors may lead to policies facilitating the integration. This could be useful in the case of emerging technologies in education. Several of these factors are categorized in [30] and are outlined below.
A category of factors involves the specific technology itself. These factors include the specific features it possesses, cost, the advantages it offers, the complexity in its use and integration, and compatibility with available infrastructure.
A second category of factors involves the educational system. A first factor is the willingness to integrate a specific technology in education as reflected in educational policies and curricula. A second one is the sufficiency of information and the existence of well-defined goals in curricula about the integration of a specific technology. A third factor is the existence of pedagogical and assessment approaches in the curricula promoting integration of a specific technology. Another factor concerns national policies ensuring the availability of corresponding hardware, learning content, and software in the educational units.
Another category of factors concerns the educational institutions. Certain factors are the technical and financial support for tutors, the size of classes, the background of students in the specific technology, the available infrastructure, and the overall practices within the institution about technology.
A further category of factors concerns tutors themselves. Several of these factors include their attitudes toward the specific technology and willingness to integrate it into education, their knowledge and experience with the specific technology, and the availability of time to learn new aspects of the technology.

3. Methodology

This research was guided by the following research questions:
  • Research Question 1: What are the main trends in the publication years, types of publication, and geographic distribution of researchers involving the work combining generative AI and robotics in K-12 and higher education?
  • Research Question 2: What are the main trends in the educational level, teaching subjects, and sample size in combinations of generative AI and robotics in K-12 and higher education?
  • Research Question 3: What are the main trends in the generative AI tools and types of robots used in combinations of generative AI and robotics in K-12 and higher education?
  • Research Question 4: How are generative AI and robotics combined in K-12 and higher education, and what are the corresponding tasks of the generative AI tools used?
  • Research Question 5: What are the main benefits and challenges reported in the approaches combining generative AI and robotics in K-12 and higher education?
The aim was to identify studies describing specific approaches that combine generative AI and robotics, rather than reviews or general overviews. To identify relevant studies, an exhaustive search was conducted in Scopus in October 2025 and February 2026 using different queries in each case. The search query was applied to the title, abstract, and keywords of the publications. This focused strategy reduced the retrieval of irrelevant studies in which the corresponding terms were only mentioned in the full text.
The search query was formulated to include relevant terms. For the educational levels (i.e., early childhood, primary, secondary, and higher education), alternative terms were included to reflect the terminology used in different countries. For early childhood education, the terms “early childhood education”, “preschool”, and “kindergarten” were included. For primary education, the terms “primary education”, “elementary education”, “primary school”, and “elementary school” were used. For secondary education, the terms “secondary education”, “middle school”, “middle education”, “high school”, and “secondary school” were included. For higher education, the terms “higher education”, “undergraduate”, “postgraduate”, “post-graduate”, “college”, and “universit*” were used. The query also included the term “ChatGPT” due to its widespread popularity and the possibility that some researchers may use it without explicitly referring to “generative AI” or “generative artificial intelligence.” In addition, the general educational terms “learner”, “teacher”, and “student” were included.
The specific query used to do the search in October 2025 is as follows:
(“generative AI” OR “generative artificial intelligence” OR ChatGPT) AND (“robot*” OR “robotic*”) AND (“kindergarten*” OR “early childhood education” OR “preschool*” OR “primary education” OR “primary school*” OR “elementary school*” OR “elementary education” OR “secondary school*” OR “secondary education” OR “high school*” OR “middle school*” OR “middle education” OR “K-12 education” OR “K-12 school*” OR “higher education” OR undergraduate OR postgraduate OR “post-graduate” OR “college*” OR “universit*” OR “student*” OR “learner*” OR “teacher*”)
The set of items retrieved in October 2025 was analyzed, and the relevant papers were studied, reaching useful conclusions for structuring the presentation of the results and the discussion in this review.
The query used for the search conducted in February 2026 included all the aforementioned terms, except for the first three, which were replaced with the terms “LLM*” and “large language model*”. This modification was made because certain researchers were found to use these terms in their publications without referring to “generative AI” or “generative artificial intelligence”. In this case, the search covered all years up to the end of 2025. The two sets of items retrieved in October 2025 and February 2026 were subsequently merged, with duplicates removed.
A total of 417 items were initially retrieved. Of these, 232 were excluded after screening their titles and abstracts. More specifically, the following items were excluded at this stage:
  • Items not accessible through the infrastructure of our institution or not generally accessible (29 items),
  • Items not written in English (2 items),
  • Editorials, conference reviews, and notes (88 items),
  • Overviews and reviews outside the subject area (40 items),
  • Studies not involving education (13 items),
  • Studies not involving robotics (38 items),
  • Books outside the subject area (14 items),
  • Studies not involving the combination of generative AI and robotics (6 items),
  • Studies not involving K-12 or higher education (2 items).
From the remaining 185 papers (i.e., those not excluded based on their titles and abstracts), 26 papers were included in the review without full-text examination, as they were clearly relevant. The full text of the remaining 159 papers was examined to determine eligibility, resulting in 131 exclusions. The following papers were excluded in this phase:
  • Studies not involving education (2 items),
  • Studies not involving K-12 or higher education (3 items),
  • Studies not involving robotics (90 items),
  • Studies not involving generative AI (5 items),
  • Items not explicitly studying the combination of generative AI and robotics (10 items),
  • Items without explicit educational use (17 items),
  • Overviews or reviews not covering the whole subject area of the review (2 items),
  • Workshop overviews (2 items).
The total number of excluded papers was 363, leaving 54 papers included in this review. Table 1 lists all included papers, along with their year of publication, publication type, and authors’ countries. For papers with co-authors from different countries or multiple affiliations, the corresponding column in Table 1 lists all relevant countries.
Table 1. List of surveyed studies.

5. Discussion and Future Research Directions

Five research questions were defined for this review, as presented in Section 3, and addressed in Section 4. More specifically, each research question is answered in Section 4.1, Section 4.2, Section 4.3, Section 4.4 and Section 4.5, respectively. This section discusses the results and highlights future research directions.
The analysis of the surveyed studies reveals several benefits deriving from the combinations of generative AI and robotics. These benefits were presented and categorized in Section 4.5 and concern both students and educators. They could form the basis when considering the implementation of relevant approaches. However, they provide only a partial view of the learning and teaching processes. A more comprehensive understanding requires considering them in conjunction with the discussion in Section 4.4, which outlines the generative AI tasks implemented in the context of these combinations. These tasks demonstrate the added value of generative AI in robotics, as many of them would be less effective, difficult, or even impossible to implement without it. Conversely, without the robotic component, the learning experience would lose key elements and differ substantially.
The analysis also identifies several challenges associated with these combinations, as presented in Section 4.5. These challenges need to be considered when designing and implementing such approaches. The reported challenges align with factors affecting the integration of technology in education outlined in Section 2.3. The reported challenges regarding the generative AI modules align with corresponding challenges outlined in Section 2.2. As discussed below, some of these challenges align with those reported by other researchers.
A positive aspect is the broad geographic distribution of researchers across four continents, although Australia and New Zealand are not represented in the reviewed studies. Researchers from Europe and some countries on other continents (e.g., the USA, China, Japan, Taiwan, and Singapore) appear particularly active in this field, which is noteworthy. However, it is likely that researchers from other regions will become increasingly active in the near future. The aforementioned findings align with results presented in [84] showing a global digital divide in generative AI adoption. Some of the factors affecting country-level generative AI usage are the digital infrastructure, English fluency, and human capital. Therefore, usage intensity remains substantially higher in high-income economies, including Europe and North America.
Most studies were implemented in higher education, although other educational levels are also represented. This could be attributed to specific reasons. First, robotics courses are an integral part of the curricula in many higher education fields. Second, higher education institutions are more likely to possess the resources required to implement such approaches, including developers, technically experienced instructors and students, servers, software, robots, and related hardware. Educational institutions at other levels may require additional support to adopt similar approaches, representing a potential direction for future research.
The sample size in most of the surveyed papers involved thirty participants or fewer. Implementations on a larger scale could assess better the corresponding approaches. Given that most approaches seem to generally be implemented in higher education, it could be feasible to present results with a larger number of participants.
Several generative AI tools were used in the studies, which were categorized in Section 4. Generally speaking, an under-explored direction is to compare the effectiveness of alternative generative AI tools in doing the same tasks in robotics. This is another future research direction. For instance, one could compare the effectiveness of various conversational AI tools (e.g., ChatGPT, Gemini, Copilot, Claude AI, and open tools) in performing corresponding tasks in robotics.
Some reviewed studies compare alternative generative AI tools. For example, in [68], the open LLMs Gemma, Llama, Mistral, oLMo, and Qwen run locally and are compared in the context of generating educational game scenarios. In [33], Gemini and two versions of GPT are thoroughly compared in interpreting young children’s instructions regarding the actions to be performed by a robotic arm. The comparison was done in terms of interpretation effectiveness and time efficiency as both parameters are important in HRI involving young students. In [56], GPT and models running locally (Llama, Mixtral, and GPT2LMHead) are compared in supporting students to learn about the functionality of an educational robot. The models were assessed based on how relevant and informative were the responses to questions.
It would be useful to have further studies comparing alternative generative AI tools in terms of effectiveness and time efficiency. The aforementioned studies compare conversational AI tools. Studies comparing tools belonging to other categories of generative AI tools (e.g., code assistance, text-to-speech, generative art, 3D modeling, and design) would be beneficial.
As far as conversational AI is concerned, tools featuring customized interfaces to GPT were found to be useful, as they were tailored to the needs of students in robotics. These tools facilitated the submission of specialized prompts (e.g., creation of web pages, robot programming, course syllabi, and robotic kit assembly) [44]. However, their implementation requires institutional support, although it may represent a valuable practice.
In [11], different types of ChatGPT use among learners in a university department are described, along with illustrative examples. These usage types support learning and are summarized in Table 8. They were selected to be applicable across the scientific domains of university departments and may therefore align with broader uses of ChatGPT (or conversational AI in general) in contexts where generative AI supports learning. Uses explicitly reported in the surveyed studies are indicated in italics in Table 8.
Table 8. Types of conversational AI use addressed to higher education students [11].
For example, students used conversational AI to retrieve information about course syllabi [44], for brainstorming and idea generation [67], and, in one study, to assist with report writing, i.e., academic writing support. It is likely that additional uses were employed but not explicitly reported (e.g., retrieving relevant examples, clarifying ambiguous or abstract theoretical concepts, and accessing tutorials). As a direction for future work, it would be valuable to systematically document the types of conversational AI use among students in robotics courses. Initial efforts in this direction have already been made in some of the surveyed studies.
Physical robots acquired from the market are the type most frequently mentioned in the surveyed papers. In most cases, these are well-known social robots. When financial resources are available, acquiring established commercial robots is a convenient option. It is noteworthy that a robotic arm and a robot cleaner were employed in two studies, indicating that a wide range of commercially available physical robots can be used in educational settings. This is further illustrated by a study utilizing a UAV simulation.
Robots developed by researchers are also used in some studies. Such robots offer flexibility in terms of functionality, hardware design, cost, and alignment with the curriculum, allowing them to address the specific needs of different educational contexts.
Most of the surveyed papers focus on two main types of GenAI-robotics integration: (i) GenAI supporting learning in robotics and (ii) emphasis on GenAI-powered robot functionality. These appear to be the most prevalent approaches. This can be explained as follows. First, approaches in which GenAI supports learning in robotics have primarily been applied in higher education, where robotics courses or course components are commonly offered and there is strong interest in enhancing learning through AI, particularly generative AI. Second, approaches emphasizing GenAI-powered robot functionality have gained attention in recent years, as generative AI helps address limitations in robot capabilities. This likely explains why this category accounts for the largest number of publications.
Approaches in which GenAI supports teaching and research in robotics appear to be relatively limited. However, it is reasonable to assume that generative AI is already being used for related tasks that have not yet been documented in the literature. For example, in [11] tasks that conversational AI (e.g., ChatGPT) can support for university instructors are analyzed. These include retrieving relevant course resources, assisting in the preparation or updating of syllabi, course outlines, lesson plans, presentations, learning materials, and assessment content. Additional applications involve retrieving and summarizing research literature and improving the presentation of academic texts. The authors suggest that faculty members likely use generative AI for such tasks, even though this usage is not explicitly reported in publications. This represents a potential direction for future research.
In the context of higher education, two additional applications of generative AI can be identified that are not discussed in [11]. The first concerns supporting educators in the design and development of robots for educational use, which can be viewed more broadly as the development of tools for teaching and learning. The second involves the use of generative AI to extract insights from students’ deliverables. These aspects could be incorporated into the discussion presented in [11].
The use of low-cost physical robots in education is an objective highlighted in certain of the surveyed studies, particularly in contexts with limited budgets. To address this, researchers have designed and developed low-cost robotic systems tailored to specific educational needs. For example, Doodlebot [79] is a social robot designed to be low-cost, reliable, and versatile. In [66], a pet-like social robot was constructed using inexpensive materials such as 3D-printed components and modeling clay. In [42], the ESP32-based embedded platform was employed in robotics activities; this platform is widely used due to its low cost and strong open-source support.
These approaches are consistent with earlier work emphasizing low-cost robotics in education, such as [90,91]. In the former, 3D printing and the ESP32 microcontroller were used to develop a low-cost robot for higher education, while in the latter, 3D printing and open-source software were used to construct a robotic manipulator. A promising direction for future research is the further exploration of low-cost robot development. In addition, proposals for scalable low-cost robotic solutions suitable for widespread adoption in education would be particularly valuable. For instance, in [79] it is suggested that Doodlebot could be deployed at scale in educational settings.
Open generative AI tools are used in ten studies, as they offer several advantages. For instance, they eliminate subscription and API metering costs and can be deployed locally. Local deployment may also reduce response latency and mitigate privacy concerns associated with proprietary tools, addressing challenges identified in Section 4.5.2.
Open language models allow researchers to fine-tune them using selected datasets tailored to specific tasks. The use of domain-specific datasets during fine-tuning can enhance performance in particular application areas [74]. This approach may also help address concerns related to the appropriateness of generated content and potential biases.
However, open tools may also present limitations. For example, in ref. [33] the authors chose not to run open-source LLMs locally due to insufficient hardware infrastructure. Furthermore, in [56] it is reported that a locally deployed open LLM without Internet access was less effective than GPT models with online access. A potential direction for future research is the broader adoption of open generative AI tools and a more systematic evaluation of their advantages and limitations.
Cost-effectiveness and scalability are key considerations in educational technology, including approaches that combine generative AI and robotics. As noted in the previous section, cost is a significant challenge. Several strategies can support cost-effective and scalable implementations. For example, locally deployed open generative AI tools eliminate subscription and API costs, offering a scalable software solution, provided that institutions can support the required infrastructure. In addition, lightweight open models with lower system requirements may offer further cost-effective and scalable options, provided they perform adequately for the intended tasks [56,68,74]. However, the limitations of open tools must also be taken into account.
Low-cost robots can further contribute to cost-effectiveness and scalability. In some studies, researchers developed low-cost robotic systems within their institutions, representing an additional viable approach. For instance, in [79] it is reported that their robot could be deployed at scale in educational settings. Moreover, as discussed in Section 2.1, robot simulations provide a cost-effective and scalable alternative, particularly in higher education. Examples are UAV simulations as exemplified in a study.
Multimodal language models are used in a limited number of studies (e.g., [43,45,51,73]). Popular generative AI tools (e.g., ChatGPT, Gemini, Copilot, Claude, and Llama) have recently incorporated multimodal capabilities, whereas earlier versions primarily supported text-based interaction. For example, a multimodal version of Llama was employed in [73]. In this context, vision-language models appear particularly promising in medical education [43,45,51]. As users become more accustomed to multimodal interaction, further studies exploring these capabilities are expected to emerge. Important aspects of multimodal interaction in educational settings are time efficiency and cost. For example, in [58] it is mentioned that the more recent version of GPT (GPT 4) supports visual, audio, and text inputs, but it can be expensive and slower for real-time applications compared to the previous version (GPT 3.5), which does not provide this multimodal functionality. This means that an examination of alternative generative AI tools may be needed along with the use of local deployment where feasible. Furthermore, the combination of generative AI tools with other tools providing visual understanding and speech processing may provide solutions [58].
Equity issues constitute another important aspect of the generative AI and robotics combinations. As noted in the Background section, a key challenge is the digital divide, which refers to inequalities in access to and use of digital technologies. Open generative AI tools and low-cost robots may help mitigate cost-related inequities. Further equity issues such as English fluency are highlighted in [26] in the context of the global digital divide in generative AI adoption. More equity issues are highlighted in [29] in the general context of the digital divide. However, additional equity concerns can be identified.
The first relates to insufficient datasets for certain languages [3], which hinders the development of effective AI models. Furthermore, in some of the reviewed multilingual studies, specific modules were found to be less effective across all languages, requiring content to be translated in order to be processed accurately. A third issue concerns difficulties in speech processing for younger users. Young children’s speech differs from that of older individuals, which can reduce the accuracy of speech-to-text systems. The resulting text is then processed by language models, which may misinterpret the intended meaning and produce inaccurate responses [33]. Notably, in [33] Gemini was found more accurate than GPT in responding to young children’s spoken instructions, highlighting the need for robust systems tailored to younger learners.
Additional equity concerns include potential biases related to gender, race, nationality, language, and culture embedded in generative AI systems. Finally, limited access to robotics and generative AI technologies remains a significant issue.
Criteria for selecting generative AI tools used in combination with robotics can be defined across three categories: (a) general criteria applicable to both proprietary and open tools, (b) criteria specific to proprietary tools, and (c) criteria specific to locally deployed open tools. In general, and based on findings from some reviewed studies, it is advisable to compare alternative tools and select the most appropriate ones. The main criteria are outlined below.
A primary general criterion concerns the tasks the tools are expected to implement. Based on these tasks, relevant tools can be identified, for example, through a review of existing studies. Closely related is the requirement for specific input and output modalities (e.g., text, images, speech, video, 3D models, code, or tabular data). Tools that support the necessary modalities should be prioritized. These criteria help identify candidate tools, whose effectiveness must then be evaluated.
Time efficiency in task execution is also important, as it can enhance learner engagement and overall interaction quality. Usability is another key consideration. In addition, tools should adequately support the language(s) used in the educational context. Finally, alignment with the broader cultural and social context (e.g., nationality, race, and cultural norms) should also be taken into account.
Specific criteria for proprietary tools include the availability of a free version, which can facilitate access for students and educators outside institutional environments. Subscription costs for both institutions and individuals should also be considered in relation to available budgets. Similarly, API metering costs and token limits should align with institutional constraints and interaction requirements [27]. Another important factor is the ability to customize or fine-tune models remotely using specialized datasets. Finally, when multiple APIs are available, selection should be based on a balance between cost and token limitations [27].
Specific criteria for open tools include hardware requirements and whether the available infrastructure can support local deployment. The extent to which models can be fine-tuned is also important, as is the ease of customization.
The combination of generative AI and robotics may be extended to other human-related fields. One of these fields is healthcare training. Two such directions are discerned in the reviewed studies. For example, some of the reviewed studies address the use of generative AI to support learning in robotic surgery [43,45,51]. As mentioned in these studies, this aspect generally concerns undergraduate students (i.e., surgery is a part of undergraduate studies) and junior doctors wanting training in advanced surgical methods. Perhaps experienced doctors not acquainted with some recent trends in robotic surgery may also benefit from such approaches. The other direction concerns the use of social robots combined with language models to simulate virtual patients [35,36,37].
Furthermore, as outlined in the Background section, various types of robots are used in medicine to perform several tasks beyond surgery. A future direction is to study combinations considering robots performing other types of tasks besides the ones focusing on surgery.
GenAI-powered robots in education may provide services to learners, educators, and institutions besides teaching, learning, and research. For example, GenAI-powered social robots may be used to navigate freshmen students and visitors in the facilities of educational institutions. Relevant approaches are discussed in [92,93,94,95]. In this context, robots need to be able to interact effectively with humans in verbal and non-verbal ways, displaying emotions, responding to their questions, and guiding them in the facilities. To achieve this, the robots need to have access to information about the educational institution and the facilities. This direction may be adapted for facilities in other sectors such as libraries, hospitals, government buildings [92], museums, and exhibition centers.
A further type of a service that GenAI-powered social robots may provide to students beyond learning and teaching involves counseling. For example, in [96], an approach using a GenAI-powered social robot is presented to assist students in selecting the vocational training program most suitable to them. Two main issues need to be dealt with in this context. The former is the large number of available vocational training programs, which creates difficulties in the selection process. The latter concerns the unwillingness of students to interact with career counselors. Career counselors may employ social robots to facilitate the counseling process since they may be appealing to students. According to the status of counselors (i.e., full-time or part-time staff members, visitors, and mobile staff members), this may be done in counseling hours, career fairs, and during school visits [96]. Similarly, GenAI-powered social robots may be used in other counseling tasks.
The search in Scopus identified a review published in conference proceedings [97] and a book chapter [76] addressing the topic of social robots integrated with generative AI in education. However, neither covers the full scope of the present review and is therefore discussed only briefly.
The review in [97] represents an early effort to examine the integration of ChatGPT and social robots in learning. As the corresponding proceedings were published in early 2024, the body of relevant literature was still limited. The authors identified ten papers, only three of which were related to education; the remaining studies focused on service domains, healthcare, and general applications. Of the three education-related papers, two focused on chatbots and one on social robots. The authors explicitly note that they were unable to identify work integrating ChatGPT and social robots in educational contexts. They provide background information.
Compared to the present review, differences can be noted. First, they focus specifically on ChatGPT rather than generative AI more broadly and therefore do not consider earlier work such as [80], which employed GAID and a social robot in early childhood education. Second, their review does not address the combination of generative AI with robots that are not social, a direction that is of interest in educational contexts and is covered in this review. Third, one of their eligibility criteria requires that studies provide information directly relevant to students (e.g., academic advice). It is unclear whether this criterion would exclude studies included in the present review, such as those examining the role of generative AI in supporting teaching and research in robotics or studies such as [70], which investigate educators’ perspectives.
The book chapter in [76] presents an overview of their vision of autonomous and adaptive social robots in education enhanced by generative AI. It discusses limitations of current social educational robots, recent advances in generative AI, applications of generative AI in educational robotics, and technical and societal limitations of generative AI. However, it is an overview rather than a review and focuses specifically on generative AI in social robots rather than the broader landscape of generative AI and robotics combinations.
Certain challenges identified in Section 4.5.2 are also mentioned in [76]. These concern the response delays, generative AI inaccuracy and hallucinations, anthropomorphism, trust and deception, and speech recognition limitations. They also mention the difficulty of operating local generative AI tools in case the background knowledge of staff members is insufficient, which aligns, to a certain degree, with the challenging infrastructure requirements mentioned in Section 4.5.2. Our review seems to have encountered further challenges. A challenge mentioned in [76] not discussed in Section 4.5.2 concerns the issue of sustainability, which means that it is good to choose options that are as sustainable as possible. Furthermore, our review discusses benefits for learners and educators based on the reviewed work (Section 4.5.1) and generative AI tasks in the context of GenAI and robotics combinations. In [76], the focus is on benefits generative AI provides to social robots, which lead to benefits for learners.
The limitations of this review should also be acknowledged. First, only the Scopus database was used to retrieve relevant studies, due to its broad coverage of journals and conference proceedings. However, additional results might have been identified by including other databases. Second, the search was limited to titles, abstracts, and keywords in order to avoid retrieving irrelevant studies that merely mention the search terms. This practice is done in other reviews as well. However, it may be possible that some more relevant work could be retrieved by searching in the full text. Moreover, work not authored in English was excluded for practical reasons. This is a typical practice in reviews but may have excluded work presenting useful results. Finally, we excluded work that was not accessible during the period the search was carried out. This is also a typical practice in reviews but may have excluded some relevant work.

6. Conclusions

This paper reviews research on the combinations of generative AI and robotics in K–12 and higher education. Most of the work has been published during the last two years. The goal of these combinations is to enhance learning and teaching experiences by leveraging the strengths of both technologies.
The selected studies are analyzed, and key trends are discussed. The identified approaches span all educational levels and a variety of teaching subjects. Generative AI tools are categorized into five groups based on their primary functions. In addition, different types of robots are employed, demonstrating the versatility of the implemented approaches.
Five main types of combinations are identified, and the roles of generative AI within each category are outlined. The categorization itself is a good step for viewing the whole perspective of GenAI and robotics combinations. It deals with multiple viewpoints on the combinations of GenAI and robotics. It covers six viewpoints: (i) teachers (how GenAI supports teaching and research in robotics), (ii) learners (how GenAI supports learning in robotics), (iii) robotics as a teaching subject taught by a GenAI-powered system, (iv) AI literacy in the context of combinations, (v) GenAI-powered robots, and (vi) consideration of various types of robots in (i)–(v). To the best of our knowledge, previous reviews and overviews do not cover all of these six viewpoints.
Several benefits for both learners and educators are highlighted, indicating the effectiveness of these approaches. As expected, a number of challenges are also reported, some of which are consistent with findings in related work.
This combination has the potential to offer significant advantages to both students and educators. In the coming years, the number of relevant studies is likely to increase. We believe that this review provides a basis for further discussion in this field.

Author Contributions

Conceptualization, J.P.; methodology, J.P.; validation, J.P. and A.B.; data curation, J.P. and A.B.; writing—original draft preparation, J.P. and A.B.; writing—review and editing, J.P. and A.B.; visualization, J.P. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Data sharing is not applicable.

Conflicts of Interest

The authors declare no conflicts of interest.

Abbreviations

The following abbreviations are used in this manuscript:
AAAIAmerican association for artificial intelligence
ACMAssociation for computing machinery
AIArtificial intelligence
AIEDArtificial intelligence in education
APIApplication programming interface
ASEEAssociation for engineering education
ASRAutomatic speech recognition
CADComputer-aided design
CAEComputer-aided engineering
CAMComputer-aided manufacturing
CSComputer science
ECEEarly childhood education
EFLEnglish as a foreign language
EREducational robotics
GAIDGenerative artificial intelligence drawing
GenAIGenerative artificial intelligence
GPTGenerative pre-trained transformer
HIEHigher education
HRIHuman-robot interaction
IEEEInstitute of electrical and electronics engineers
LLMLarge language model
PEPrimary education
ROSRobot Operating System
SESecondary education
STEMScience, technology, engineering, and mathematics
UAVUnmanned aerial vehicle

Appendix A

Table A1 summarizes the benefits and challenges reported in the surveyed studies. The acronyms ‘CS’, ‘AIED’, ‘ASR’, ‘ER’, and ‘ROS’ stand for computer science, AI in education, automatic speech recognition, educational robotics, and Robot Operating System, respectively.
Table A1. Benefits and challenges reported in the surveyed studies.

References

  1. Hughes, J.E.; Roblyer, M.D. Integrating Educational Technology into Teaching: Transforming Learning Across Disciplines, 9th ed.; Pearson: New York, NY, USA, 2023. [Google Scholar]
  2. Prentzas, J. Artificial Intelligence methods in early childhood education. In Artificial Intelligence, Evolutionary Computing and Metaheuristics. In the Footsteps of Alan Turing; Yang, X.S., Ed.; Studies in Computational, Intelligence; Springer: Berlin/Heidelberg, Germany, 2013; Volume 427, pp. 169–199. [Google Scholar] [CrossRef] [Scilit]
  3. Prentzas, J.; Binopoulou, A. Explainable artificial intelligence approaches in primary education: A review. Electronics 2025, 14, 2279. [Google Scholar] [CrossRef] [Scilit]
  4. Martin, F.; Zhuang, M.; Schaefer, D. Systematic review of research on artificial intelligence in K-12 education (2017–2022). Comput. Educ. Artif. Intell. 2024, 6, 100195. [Google Scholar] [CrossRef] [Scilit]
  5. Zawacki-Richter, O.; Marín, V.I.; Bond, M.; Gouverneur, F. Systematic review of research on artificial intelligence applications in higher education—Where are the educators? Int. J. Educ. Technol. High Educ. 2019, 16, 39. [Google Scholar] [CrossRef] [Scilit]
  6. Chen, L.; Chen, P.; Lin, Z. Artificial intelligence in education: A review. IEEE Access 2020, 8, 75264–75278. [Google Scholar] [CrossRef] [Scilit]
  7. Perikos, I.; Grivokostopoulou, F.; Hatzilygeroudis, I. Assistance and feedback mechanism in an Intelligent Tutoring System for teaching conversion of Natural Language into Logic. Int. J. Artif. Intell. Educ. 2017, 27, 475–514. [Google Scholar] [CrossRef] [Scilit]
  8. Chrysafiadi, K.; Virvou, M. Evaluating the integration of fuzzy logic into the student model of a web-based learning environment. Expert Syst. Appl. 2012, 39, 13127–13134. [Google Scholar] [CrossRef] [Scilit]
  9. Ahmad, K.; Iqbal, W.; El-Hassan, A.; Qadir, J.; Benhaddou, D.; Ayyash, M.; Al-Fuqaha, A. Data-driven artificial intelligence in education: A comprehensive review. IEEE Trans. Learn. Technol. 2023, 17, 12–31. [Google Scholar] [CrossRef] [Scilit]
  10. Prentzas, J.; Sidiropoulou, M. Assessing the use of Open AI Chat-GPT in a University Department of Education. In Proceedings of the 14th International Conference on Information, Intelligence, Systems and Applications (IISA 2023), Volos, Greece, 10–12 July 2023; IEEE: New York, NY, USA, 2023; pp. 1–4. [Google Scholar] [CrossRef] [Scilit]
  11. Prentzas, J.; Sidiropoulou, M. Integrating OpenAI Chat-GPT in a University Department of Education: Main types of use and preliminary assessment results. In Extended Selected Papers of the 14th International Conference on Information, Intelligence, Systems, and Applications; Bourbakis, N., Tsihrintzis, G.A., Virvou, M., Jain, L.C., Eds.; Lecture Notes in Networks and Systems; Springer: Cham, Switzerland, 2024; Volume 1093, pp. 306–326. [Google Scholar]
  12. Ben-Ari, M.; Mondada, F. Elements of Robotics; Springer Nature: Cham, Switzerland, 2018. [Google Scholar]
  13. Ryalat, M.; Almtireen, N.; Al-refai, G.; Elmoaqet, H.; Rawashdeh, N. Research and education in robotics: A comprehensive review, trends, challenges, and future directions. J. Sens. Actuator Netw. 2025, 14, 76. [Google Scholar] [CrossRef] [Scilit]
  14. Adorni, G.; Grosso, D. Speeding up design and making to reduce time-to-project and time-to-market: An AI-Enhanced approach in engineering education. In Proceedings of the Second International Workshop on Artificial Intelligent Systems in Education (AIxEDU 2024) Co-Located with the 23rd International Conference of the Italian Association for Artificial Intelligence (AIxIA 2024), Bolzano, Italy, 26 November 2024; CEUR Workshop Proceedings. Volume 3879. [Google Scholar]
  15. Dong, S.; Zhi, R.; Gan, F. Integrating robotics in art education: Leveraging constructivist and experiential learning frameworks to enhance cognitive development, problem-solving skills and collaboration among diverse learners. Eur. J. Educ. 2025, 60, e70114. [Google Scholar] [CrossRef] [Scilit]
  16. Cangelosi, A.; Stramandinoli, F. A review of abstract concept learning in embodied agents and robots. Philos. Trans. R. Soc. B Biol. Sci. 2018, 373, 20170131. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  17. Breazeal, C.; Dautenhahn, K.; Kanda, T. Social robotics. In Springer Handbook of Robotics; Siciliano, B., Khatib, O., Eds.; Springer: Cham, Switzerland, 2016; pp. 1935–1972. [Google Scholar] [CrossRef] [Scilit]
  18. Mubin, O.; Stevens, C.J.; Shahid, S.; Al Mahmud, A.; Dong, J.J. A review of the applicability of robots in education. J. Technol. Educ. Learn. 2013, 1, 13. [Google Scholar] [CrossRef] [Scilit]
  19. Verhelst, E.; Janssens, R.; Belpaeme, T. Enabling autonomous and adaptive social robots in education: A vision for the application of Generative AI. In Social Robots in Education; Lampropoulos, G., Papadakis, S., Eds.; Studies in Computational Intelligence; Springer: Cham, Switzerland, 2025; Volume 1194, pp. 17–42. [Google Scholar] [CrossRef] [Scilit]
  20. Dupont, P.E.; Nelson, B.J.; Goldfarb, M.; Hannaford, B.; Menciassi, A.; O’Malley, M.K.; Simaan, N.; Valdastri, P.; Yang, G.Z. A decade retrospective of medical robotics research from 2010 to 2020. Sci. Robot. 2021, 6, eabi8017. [Google Scholar] [CrossRef] [Scilit]
  21. Trapero-González, I.; Hinojo-Lucena, F.J.; Romero-Rodríguez, J.M.; Martínez-Menéndez, A. Didactic impact of educational robotics on the development of STEM competence in primary education: A systematic review and meta-analysis. Front. Educ. 2024, 9, 1480908. [Google Scholar] [CrossRef] [Scilit]
  22. Xia, L.; Zhong, B. A systematic review on teaching and learning robotics content knowledge in K-12. Comput. Educ. 2018, 127, 267–282. [Google Scholar] [CrossRef] [Scilit]
  23. Zamora, P.; Lozada, A.; Buele, J.; Avilés-Castillo, F. Robotics in higher education and its impact on digital learning. Front. Comput. Sci. 2025, 7, 1607766. [Google Scholar] [CrossRef] [Scilit]
  24. Ztoupas, E.; Sapounidis, T.; Tselegkaridis, S. Simulators in educational robotics: A systematic review with content analysis. Appl. Sci. 2026, 16, 653. [Google Scholar] [CrossRef] [Scilit]
  25. Gozalo-Brizuela, R.; Garrido-Merchán, E.C. A survey of generative AI applications. arXiv 2023, arXiv:2306.02781. [Google Scholar] [CrossRef] [Scilit]
  26. Liu, Y.; Wang, H. Who on earth is using Generative AI? World Dev. 2026, 199, 107260. [Google Scholar] [CrossRef] [Scilit]
  27. Tozadore, D.; Rusu, A.-M. Teachers, take care of the essential. the rest is story: Using LLM and social robots for content approaching by storytelling. In Proceedings of the 33rd IEEE International Workshop on Robot and Human Communication (RO-MAN 2024), Pasadena, CA, USA, 26–30 August 2024; IEEE: New York, NY, USA, 2024; pp. 2125–2130. [Google Scholar] [CrossRef] [Scilit]
  28. Nah, F.F.-H.; Zheng, R.; Cai, J.; Siau, K.; Chen, L. Generative AI and ChatGPT: Applications, challenges, and AI-human collaboration. J. Inf. Technol. Case Appl. Res. 2023, 25, 277–304. [Google Scholar]
  29. Lythreatis, S.; Singh, S.K.; El-Kassar, A.N. The digital divide: A review and future research agenda. Technol. Forecast. Soc. Change 2022, 175, 121359. [Google Scholar] [CrossRef] [Scilit]
  30. Karipidis, N.; Prentzas, J. A survey of factors affecting the successful integration of ICT in education. In Proceedings of the 10th International Technology, Education and Development Conference (INTED 2016), Valencia, Spain, 7–9 March 2016; International Academy of Technology, Education and Development: Valencia, Spain, 2016; pp. 8456–8466. [Google Scholar] [CrossRef] [Scilit]
  31. Abbas, I.; Ganesan, S. Exploring the feasibility of AI-driven educational robots in simulation environments for early childhood learning: A pilot study. In Proceedings of the IEEE International Conference on Electro Information Technology (eIT 2025), Valparaiso, IN, USA, 29–31 May 2025; IEEE: New York, NY, USA, 2025; pp. 173–181. [Google Scholar] [CrossRef] [Scilit]
  32. Abbo, G.A.; Desideri, G.; Belpaeme, T.; Spitale, M. ‘Can you be my mum?’: Manipulating social robots in the large language models era. In Proceedings of the 20th ACM/IEEE International Conference on Human-Robot Interaction, Melbourne, Australia, 4–6 March 2025; IEEE: New York, NY, USA, 2025; pp. 1181–1185. [Google Scholar] [CrossRef] [Scilit]
  33. Aguilera, C.A.; Castro, A.; Aguilera, C.; Rǎducanu, B. Voice-controlled robotics in early education: Implementing and validating child-directed interactions using a collaborative robot and Artificial Intelligence. Appl. Sci. 2024, 14, 2408. [Google Scholar] [CrossRef] [Scilit]
  34. Ashok, A.; Bruno, B.; Helf, T.; Berns, K. ‘Thanks for the Practice!’: LLM-powered social robot as Tandem language partner at university. In Proceedings of the 20th ACM/IEEE International Conference on Human-Robot Interaction (HRI 2025), Melbourne, Australia, 4–6 March 2025; IEEE: New York, NY, USA, 2025; pp. 1221–1226. [Google Scholar] [CrossRef] [Scilit]
  35. Borg, A.; Jobs, B.; Huss, V.; Gentline, C.; Espinosa, F.; Ruiz, M.; Edelbring, S.; Georg, C.; Skantze, G.; Parodis, I. Enhancing clinical reasoning skills for medical students: A qualitative comparison of LLM-powered social robotic versus computer-based virtual patients within rheumatology. Rheumatol. Int. 2024, 44, 3041–3051. [Google Scholar] [CrossRef] [Scilit]
  36. Borg, A.; Georg, C.; Jobs, B.; Huss, V.; Waldenlind, K.; Ruiz, M.; Edelbring, S.; Skantze, G.; Parodis, I. Virtual patient simulations using social robotics combined with large language models for clinical reasoning training in medical education: Mixed methods study. J. Med. Internet Res. 2025, 27, e63312. [Google Scholar] [CrossRef] [Scilit]
  37. Borg, A.; Schiött, J.; Ivegren, W.; Gentline, C.; Huss, V.; Hugelius, A.; Jobs, B.; Espinosa, F.; Ruiz, M.; Edelbring, S.; et al. AI-enhanced social robotic versus computer-based virtual patients for clinical reasoning training in medical education: Observational crossover cohort study. J. Med. Internet Res. 2025, 27, e82541. [Google Scholar] [CrossRef] [Scilit]
  38. Brender, J.; El-Hamamsy, L.; Mondada, F.; Bumbacher, E.W. Who’s helping who? When students use ChatGPT to engage in practice lab sessions. In Artificial Intelligence in Education; Olney, A.M., Chounta, I.A., Liu, Z., Santos, O.C., Bittencourt, I.I., Eds.; Lecture Notes in Computer Science; Springer: Cham, Switzerland, 2024; Volume 14829, pp. 235–249. [Google Scholar] [CrossRef] [Scilit]
  39. Brender, J.; El-Hamamsy, L.; Uittenhove, K.; Mondada, F.; Bumbacher, E.W. Structured prompts, better outcomes? Exploring the effects of a structured interface with ChatGPT in a graduate robotics course. In Proceedings of the 20th European Conference on Technology Enhanced Learning (EC-TEL 2025), Durham, UK, 15–19 September 2025; pp. 91–106. [Google Scholar] [CrossRef] [Scilit]
  40. Buchem, I.; Bäcker, N.; Trutty, A.; Thomas, E.; Dincel, K. Humour in educational robots: Investigating the effects of humour in a robot-led Scrumban simulation in business education. In Proceedings of the 16th International Conference on Computer Supported Education (CSEDU 2024), Angers, France, 2–4 May 2024; SciTePress: Setúbal, Portugal, 2024; Volume 1, pp. 314–321. [Google Scholar] [CrossRef] [Scilit]
  41. Chakraborty, S.; Loyens, D.; Aston, J.H. Product ideation in the age of Artificial Intelligence: Insights on design process through shape coding social robots. Open Educ. Stud. 2025, 7, 20250094. [Google Scholar] [CrossRef] [Scilit]
  42. Chen, H.; Luo, J.; Shi, W.; Wu, J.; Wang, J.; Leach, M.P.; Zhang, X.; Zhang, Q. Development of an AI desktop assistant based on ESP32: Practice on experiential learning in engineering education. In Proceedings of the 14th International Conference on Information Technology in Medicine and Education (ITME 2024), Guiyang, China, 13–15 September 2024; IEEE: New York, NY, USA, 2024; pp. 1156–1160. [Google Scholar] [CrossRef] [Scilit]
  43. Chen, K.; Du, Y.; You, T.; Hoque, M.; Guo, Z.; Jin, Y.; Chen, G.; Heng, P.A. LLM-Assisted Multi-Teacher Continual Learning for Visual Question Answering in Robotic Surgery. In Proceedings of the IEEE International Conference on Robotics and Automation (ICRA 2024), Yokohama, Japan, 13–17 May 2024; IEEE: New York, NY, USA, 2024; pp. 10772–10778. [Google Scholar] [CrossRef] [Scilit]
  44. Danahy, E.E.; Vora, M.K.; Xu, Y.M.; Church, W.J. Supporting students’ educational robotics experiences through Generative AI chatbots. In Proceedings of the 15th Annual First-Year Engineering Experience Conference (FYEE 2024), Boston, MA, USA, 28–30 July 2024; American Society for Engineering Education: Washington, DC, USA, 2024; p. 45077. [Google Scholar]
  45. Du, Y.; Chen, K.; Zhan, Y.; Low, C.H.; Hoque, M.; Guo, Z.; Jin, Y.; Chen, G.; Heng, P.A. LMT++: Adaptively collaborating LLMs with multi-specialized teachers for continual VQA in robotic surgical videos. IEEE Trans. Med. Imaging 2025, 44, 4678–4689. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  46. Göhler, V. Critical or confident? AI literacy and student-AI collaboration in higher education. In Proceedings of the 26th ACM Annual Conference on Cybersecurity and Information Technology Education (ACM SIGCITE 2025), Sacramento, CA, USA, 6–8 November 2025; ACM: New York, NY, USA, 2025; pp. 119–126. [Google Scholar] [CrossRef] [Scilit]
  47. González-Santamarta, M.Á.; Rodríguez-Lera, F.J.; Conde-González, M.Á.; Rodríguez-Sedano, F.; Fernández-Llamas, C. Exploring the use of LLMs for teaching AI and robotics concepts at a Master’s Degree. In Proceedings of the 20th Conference of the Spanish Association for Artificial Intelligence (CAEPIA 2024), A Coruña, Spain, 19–21 June 2024; Lecture Notes in Computer Science; Springer: Cham, Switzerland, 2024; Volume 14640, pp. 254–263. [Google Scholar] [CrossRef] [Scilit]
  48. Guntupalli, K.M.; Raja, A. FLY-LLM Sim: A novel integration of UAV and LLM lab platform. In Proceedings of the 2025 IEEE 6th Annual World AI IoT Congress, (AIIoT 2025), Seattle, WA, USA, 28–30 May 2025; IEEE: New York, NY, USA, 2025; pp. 259–265. [Google Scholar] [CrossRef] [Scilit]
  49. Hakim, V.G.A.; Yang, S.; Wang, J.; Chang, Y.; Lin, H.; Chen, G. Exploring the impact of designing a robot as a pet with Interdependence Theory on long-term relationships and learning performance. In Proceedings of the International Conference on Computers in Education, Matsue Shimane, Japan, 4–8 December 2023; Asia-Pacific Society for Computers in Education: Taoyuan City, Taiwan; pp. 611–620. [Google Scholar] [CrossRef] [Scilit]
  50. Hakim, V.G.A.; Yang, S.-H.; Wang, J.-H.; Peng, C.-Y.; Chen, G.-D. AI-powered personalized robots with digital companions: Leveraging LLMs for frequent engagement, ownership, and sustained learning. In Proceedings of the 25th IEEE International Conference on Advanced Learning Technologies (ICALT 2025), Changhua, Taiwan, 14–17 July 2025; IEEE: New York, NY, USA, 2025; pp. 54–58. [Google Scholar] [CrossRef] [Scilit]
  51. Hao, P.; Wang, H.; Yang, G.; Zhu, L. Enhancing visual reasoning with LLM-powered knowledge graphs for visual question localized-answering in robotic surgery. IEEE J. Biomed. Health Inform. 2025, 29, 9027–9040. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  52. Hasko, R.T.; Hasko, O.; Kutucu, H. Teaching assistant robots in various fields: Natural sciences, medicine and specific non-deterministic conditions. In Proceedings of the 6th International Conference on Informatics & Data-Driven Medicine (IDDM 2023), Bratislava, Slovakia, 17–19 November 2023; CEUR Workshop Proceedings. Volume 3879, pp. 303–309. [Google Scholar]
  53. Helal, M.E.; Holthaus, P.; Wood, L.J.; Velmurugan, V.; Lakatos, G.; Moros, S.; Amirabdollahian, F. When the robotic maths tutor is wrong—Can children identify mistakes generated by ChatGPT? In Proceedings of the 5th International Conference on Artificial Intelligence, Robotics and Control (AIRC 2024), Cairo, Egypt, 22–24 April 2024; IEEE: New York, NY, USA, 2024; pp. 83–90. [Google Scholar] [CrossRef] [Scilit]
  54. Hsu, T.; Hsu, T. Teaching AI with games: The impact of generative AI drawing on computational thinking skills. Educ. Inf. Technol. 2025, 30, 21499–21518. [Google Scholar] [CrossRef] [Scilit]
  55. Ito, S.; Kochigami, K.; Kanda, T. A robot dynamically asking questions in university classes. In Proceedings of the 20th ACM/IEEE International Conference on Human-Robot Interaction, Melbourne, Australia, 4–6 March 2025; IEEE: New York, NY, USA, 2025; pp. 839–848. [Google Scholar] [CrossRef] [Scilit]
  56. Krupas, M.; Antonets, L.; Vaščák, J.; Zolotova, I. AI-based assistant: LLMs for effective robotics education and research. In Proceedings of the 22nd International Conference on Emerging eLearning Technologies and Applications (ICETA 2024), Stary Smokovec, Slovakia, 24–25 October 2024; IEEE: New York, NY, USA, 2024; pp. 372–377. [Google Scholar] [CrossRef] [Scilit]
  57. Kuo, J.Y.; Xia, Z. Exploring intelligent user interfaces from design students’ perspectives on smart home products through peer assessment, focus group and ChatGPT. In Proceedings of the 30th ISTE International Conference on Transdisciplinary Engineering, Hua Hin Cha Am, Thailand, 11–14 July 2023; IOS Press: Amsterdam, The Netherlands; pp. 161–170. [Google Scholar] [CrossRef] [Scilit]
  58. Latif, E.; Parasuraman, R.; Zhai, X. PhysicsAssistant: An LLM-powered interactive learning robot for Physics lab investigations. In Proceedings of the 33rd IEEE International Workshop on Robot and Human Communication (RO-MAN 2024), Pasadena, CA, USA, 26–30 August 2024; IEEE: New York, NY, USA, 2024; pp. 864–871. [Google Scholar] [CrossRef] [Scilit]
  59. Markos, A.; Prentzas, J.; Sidiropoulou, M. Pre-service teachers’ assessment of ChatGPT’s utility in higher education: SWOT and content analysis. Electronics 2024, 13, 1985. [Google Scholar] [CrossRef] [Scilit]
  60. Mauliana, M.; Ashok, A.; Czernochowski, D.; Berns, K. Exploring LLM-powered multi-session human-robot interactions with university students. Front. Robot. AI 2025, 12, 1585589. [Google Scholar] [CrossRef] [Scilit]
  61. Narayan, M.; Saharan, L.K. Reshaping engineering technology education: Fostering critical thinking through open-ended problems in the era of Generative AI. In Proceedings of the ASEE Annual Conference & Exposition, Portland, OR, USA, 23–26 June 2024; American Society for Engineering Education: Washington, DC, USA, 2024; p. 44232. [Google Scholar]
  62. Pastore, S. ‘The Classroom Robot’, by ChatGPT: Unpacking the imaginary of the AIED classroom through an AI-generated story. Postdigit. Sci. Educ. 2025, 7, 699–718. [Google Scholar] [CrossRef] [Scilit]
  63. Pramanta, D.; Tamukoh, H. A feasibility study of generative AI applications using EV-3 Robots at the Kyushu Institute of Information Sciences. In Proceedings of the International Conference on Artificial Life and Robotics (ICAROB 2025), Oita, Japan, 13–16 February 2025; ALife Robotics Corporation Ltd.: Oita, Japan, 2025; pp. 499–502. [Google Scholar] [CrossRef] [Scilit]
  64. Puttajanyawong, N.; Suamuang, W.; Chomsuwan, K. Utilizing Generative AI to develop programming skill through self-directed and interactive learning. In Proceedings of the IEEE Global Engineering Education Conference (EDUCON 2025), London, UK, 22–25 April 2025; IEEE: New York, NY, USA, 2025; pp. 1–9. [Google Scholar] [CrossRef] [Scilit]
  65. Reeping, D.P.; Shah, A. Work-in-progress: Students’ prompting strategies when solving an engineering design task. In Proceedings of the Frontiers in Education Conference (FIE 2024), Washington, DC, USA, 13–16 October 2024; IEEE: New York, NY, USA, 2024; pp. 1–5. [Google Scholar] [CrossRef] [Scilit]
  66. Rivadeneira, F.; Carcausto, D.; Ore, C.; Saga, G.; Vilca, M.; Arroyo, D. AI—Driven design of a robotic companion with Feline-inspired behaviors for stress relief. In Proceedings of the 20th ACM/IEEE International Conference on Human-Robot Interaction (HRI 2025), Melbourne, Australia, 4–6 March 2025; IEEE: New York, NY, USA, 2025; pp. 1568–1572. [Google Scholar] [CrossRef] [Scilit]
  67. Romanillos, E.J.; Andersson, T. The intersection of AI and Bloom’s Taxonomy in design education: A robotic design case study. In Proceedings of the 26th International Conference on Engineering and Product Design Education (E&PDE 2024), Birmingham, UK, 5–6 September 2024; Department of Management and Engineering, Linköping University: Linköping, Sweden, 2024; pp. 651–656. [Google Scholar]
  68. Ruskov, M. Our coding adventure: Using LLMs to personalise the narrative of a tangible programming robot for preschoolers. In Proceedings of the Workshop on Transformative Curriculum Design—Digitalisation, Sustainability, and AI Literacy for 21st Century Learning (D-SAIL 2025), Palermo, Italy, 22 July 2025; CEUR Workshop Proceedings. 2025; Volume 4051, pp. 50–58. [Google Scholar]
  69. Rutatola, E.P.; Stroeken, K.; Belpaeme, T. “Habari, Colleague!”: A qualitative exploration of the perceptions of primary school mathematics teachers in Tanzania regarding the use of social robots. Appl. Sci. 2025, 15, 8483. [Google Scholar] [CrossRef] [Scilit]
  70. Sacco, F.; Rossini, G.; Manzi, F.; Di Dio, C.; Aquilino, L.; Cangelosi, A.; Raggioli, L.; Massaro, D.; Marchetti, A. An antropomorphic robot with ChatGPT for learning activities: The teachers’ perspective. In Proceedings of the IEEE International Conference on Metrology for eXtended Reality, Artificial Intelligence and Neural Engineering (MetroXRAINE 2023), Milano, Italy, 25–27 October 2023; IEEE: New York, NY, USA, 2023; pp. 1166–1170. [Google Scholar] [CrossRef] [Scilit]
  71. Salem, A.; Sumi, K. Deception detection in educational AI: Challenges for Japanese middle school students in interacting with generative AI robots. Front. Artif. Intell. 2024, 7, 1493348. [Google Scholar] [CrossRef] [Scilit] [PubMed]
  72. Sievers, T. A practical approach to child-robot interaction in the classroom. In Proceedings of the AAAI Symposium Series, San Francisco, CA, USA, 31 March–2 April 2025; Volume 5, pp. 425–429. [Google Scholar] [CrossRef] [Scilit]
  73. Sun, F.; Li, L.; Meng, S.; Teng, X.; Payne, T.R.; Craig, P. Integrating emotional intelligence, memory architecture, and gestures to achieve empathetic humanoid robot interaction in an educational setting. Front. Robot. AI 2025, 12, 1635419. [Google Scholar] [CrossRef] [Scilit]
  74. Tozadore, D.; Ertug, N.; Chaker, Y.; Abderrahim, M. RoboBuddy in the classroom: Exploring LLM-powered social robots for storytelling in learning and integration activities. In Proceedings of the 34th IEEE International Workshop on Robot and Human Communication (RO-MAN 2025), Eindhoven, The Netherlands, 25–29 August 2025; IEEE: New York, NY, USA, 2025; pp. 1543–1549. [Google Scholar] [CrossRef] [Scilit]
  75. Uchida, T.; Ban, M.; Koshino, A.; Maeda, M.; Sakai, K.; Matsuura, N.; Ishiguro, H.; Yoshikawa, Y. Affirmative large language model-based dialogue robots for broadening or deepening the perspective of children. Int. J. Soc. Robot. 2025, 17, 2729–2741. [Google Scholar] [CrossRef] [Scilit]
  76. Verhelst, E.; Belpaeme, T. Large language models cover for speech recognition mistakes: Evaluating conversational AI for second language learners. In Proceedings of the 20th ACM/IEEE International Conference on Human-Robot Interaction, Melbourne, Australia, 4–6 March 2025; IEEE: New York, NY, USA, 2025; pp. 1705–1709. [Google Scholar] [CrossRef] [Scilit]
  77. Verhelst, E.; Demeester, T.; Janssens, R.; Belpaeme, T. Adaptive second language tutoring using Generative AI and a social robot. In Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI 2024), Boulder, CO, USA, 11–14 March 2024; IEEE: New York, NY, USA, 2024; pp. 1080–1084. [Google Scholar] [CrossRef] [Scilit]
  78. Voutyrakou, D.A.; Skordoulis, C.D. Using AI tools to enhance educational robotics to bridge the gender gap in STEM. Educ. Sci. 2025, 15, 711. [Google Scholar] [CrossRef] [Scilit]
  79. Williams, R.; Ali, S.A.; Alcantara, R.; Burghleh, T.; Alghowinem, S.; Breazeal, C.L. Doodlebot: An educational robot for creativity and AI literacy. In Proceedings of the ACM/IEEE International Conference on Human-Robot Interaction (HRI 2024), Boulder, CO, USA, 11–14 March 2024; IEEE: New York, NY, USA, 2024; pp. 772–780. [Google Scholar] [CrossRef] [Scilit]
  80. Williams, R.; Park, H.; Breazeal, C.L. A is for artificial intelligence the impact of artificial intelligence activities on young children’s perceptions of robots. In Proceedings of the Conference on Human Factors in Computing Systems (CHI 2019), Glasgow, UK, 4–9 May 2019; ACM: New York, NY, USA, 2019; 447, pp. 1–11. [Google Scholar] [CrossRef] [Scilit]
  81. Xu, Y.M.; Danahy, E.E.; Church, W.J. Re-Design introductory engineering course for tinkering with Generative AI and the shifts in students’ perceptions of using AI for learning. In Proceedings of the ASEE Annual Conference & Exposition, Portland, OR, USA, 23–26 June 2024; American Society for Engineering Education: Washington, DC, USA, 2024; p. 43357. [Google Scholar] [CrossRef] [Scilit]
  82. Zhou, C. Application research of teaching robot based on ChatGPT. In Proceedings of the IEEE International Conference on Sensors, Electronics and Computer Engineering (ICSECE 2023), Jinzhou, China, 18–20 August 2023; IEEE: New York, NY, USA, 2023; pp. 633–637. [Google Scholar] [CrossRef] [Scilit]
  83. Bick, A.; Blandin, A.; Deming, D.J. The rapid adoption of generative AI. Manag. Sci. 2026. [Google Scholar] [CrossRef] [Scilit]
  84. Liu, J.; He, Y.; Zheng, Z.; Bu, Y.; Ni, C. AI-assisted writing is growing fastest among non-English-speaking and less established scientists. arXiv 2025, arXiv:2511.15872. [Google Scholar]
  85. Bai, Y.; Kadavath, S.; Kundu, S.; Askell, A.; Kernion, J.; Jones, A.; Chen, A.; Goldie, A.; Mirhoseini, A.; McKinnon, C.; et al. Constitutional AI: Harmlessness from AI feedback. arXiv 2022, arXiv:2212.08073. [Google Scholar] [CrossRef] [Scilit]
  86. Wermelinger, M. Using Github Copilot to solve simple programming problems. In Proceedings of the 54th ACM Technical Symposium on Computer Science Education (SIGCSE 2023), Toronto, Canada, 15–18 March 2023; ACM: New York, NY, USA, 2023; pp. 172–178. [Google Scholar] [CrossRef] [Scilit]
  87. Moniz, A.; Gordon, M.; Bergum, I.; Chang, M.; Grant, G. Beginning Azure Cognitive Services; Apress: Berkeley, CA, USA, 2021. [Google Scholar]
  88. Bhad, P.U.; Buktar, R.B. Multiple novel generative design solutions for various mechanical engineering related products using Autodesk Fusion 360 software. Int. J. Des. Eng. 2022, 11, 1–26. [Google Scholar]
  89. Gaeta, A.; Orciuoli, F.; Pascuzzo, A.; Peduto, A. Enhancing traditional ITS architectures with large language models for generating motivational feedback. Comput. Educ. Artif. Intell. 2025, 9, 100433. [Google Scholar] [CrossRef] [Scilit]
  90. Ananias, E.; Gaspar, P.D. A low-cost collaborative robot for science and education purposes to foster the industry 4.0 implementation. Appl. Syst. Innov. 2022, 5, 72. [Google Scholar] [CrossRef] [Scilit]
  91. Zajc, L.Č.; Rezelj, A.; Skočaj, D. Low-cost open-source robotic platform for education. IEEE Trans. Learn. Technol. 2022, 16, 18–25. [Google Scholar] [CrossRef] [Scilit]
  92. Aldini, M.O.A.; Ismail, A.H.; Ayob, M.N.; Basri, H.H.; Wen, L.J.; Thongkot, D. Development of DeepSeek-based conversational robot system for indoor navigation assistance in FKTE1 Building, UniMAP. In Proceedings of the 9th International Conference on Man-Machine Systems (ICoMMS 2025), Malacca, Malaysia, 18–19 August 2025; IEEE: New York, NY, USA, 2025; pp. 198–203. [Google Scholar] [CrossRef] [Scilit]
  93. Jackson, I.S.; Ballard, A.G.; Hefeida, M.; Srivastava, A.; Gyawali, P. Development of microcontroller-based AI robot tour guide utilizing custom language models. In Proceedings of the IEEE 6th Annual World AI IoT Congress (AIIoT 2025), Seattle, WA, USA, 28–30 May 2025; IEEE: New York, NY, USA, 2025; pp. 1–8. [Google Scholar] [CrossRef] [Scilit]
  94. Alcantara, C.B.; Júnior, A.O.; Costa, J.; Ricken, L.; Mendes, A.; Leitão, P. Artificial Intelligence-based user interaction module for autonomous mobile service robots. In Proceedings of the IEEE 3rd Industrial Electronics Society Annual On-Line Conference (ONCON 2024), Beijing, China, 8–10 December 2024; IEEE: New York, NY, USA, 2024. [Google Scholar] [CrossRef] [Scilit]
  95. Deng, X.; Yang, D.; Zhang, Y. Intelligent orientation robot based on large language models and retrieval-augmented generation. In Proceedings of the 5th International Conference on Artificial Intelligence and Computer Engineering (ICAICE 2024), Wuhu, China, 8–10 November 2024; IEEE: New York, NY, USA, 2024; pp. 779–782. [Google Scholar] [CrossRef] [Scilit]
  96. Piazza, A.; Schacht, S.; Herzog, M. “which vocational training program is best for me”—Design of a recommender system for school students using large language models. In Proceedings of the Adjunct Proceedings of the 33rd ACM Conference on User Modeling, Adaptation and Personalization (UMAP 2025), New York, NY, USA, 16–19 June 2025; ACM: New York, NY, USA, 2025; pp. 425–428. [Google Scholar] [CrossRef] [Scilit]
  97. Jamil, N.; Saghir, F.N.; Alshanqiti, H.; Ali Almansoori, A.K.; Saeed, A.; Ahmad, A.; Belkacem, A.N. On combining the potential of social robots and ChatGPT for enhanced learning. In Proceedings of the 12th International Conference on Information and Education Technology (ICIET 2024), Yamaguchi, Japan, 18–20 March 2024; IEEE: New York, NY, USA, 2024; pp. 226–231. [Google Scholar] [CrossRef] [Scilit]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Article Metrics

Citations

Article Access Statistics

Journal Statistics
Multiple requests from the same IP address are counted as one view.
Head-Movement-Robust Micro-Expression Detection Method via 3D Motion Correction and Transformers
Previous
Research on Distracted and Fatigue-Related Driving Behavior Detection Based on YOLOv12-LAD
Next