How AI is Revolutionizing STEM Education: A New Era of Learning

With the rapid advancement of artificial intelligence (AI) technology, profound changes are occurring in the field of education. The integration of AI is set to offer new possibilities for personalized learning, intelligent tutoring, and optimized allocation of teaching resources. The application of AI in STEM education, known as AI-STEM, is emerging as a new field. As globalization and the knowledge economy develop, educational modernization becomes a key factor for countries to enhance their international competitiveness.

#### Background and Challenges of STEM Education

STEM stands for Science, Technology, Engineering, and Mathematics. STEM education is an interdisciplinary approach to learning that integrates these four disciplines, emphasizing a cross-disciplinary, integrated way of teaching. It is considered an important avenue for fostering innovative talents and addressing future challenges.

However, despite the high regard for STEM education, there are common misconceptions among educators and learners. For instance, many teachers and students equate STEM education solely with hands-on practice or maker education, or they think it is synonymous with information technology education or programming education. Others categorize STEM education as student self-exploration or collaborative inquiry, and some schools even view it as an elite education.

Last year, the UNESCO International Research Institute for STEM Education was established in Shanghai. As the 10th category 1 institute of UNESCO, its primary function is to promote inclusive, equitable, and quality education in science, technology, engineering, and mathematics from early childhood to adulthood. The institute serves as an information exchange center, network center, resource center, and capacity-building center in the field of STEM education, supporting UNESCO's strategies and the needs of its member countries, contributing to the United Nations' sustainable development agenda and world peace and development. Currently, we are actively exploring the value and applications of AI in the field of education to drive innovation and development in STEM education.

#### The Application of AI in STEM Education

Artificial intelligence is a discipline that studies how to enable computers to achieve human intelligence behaviors, including perception, reasoning, learning, and decision-making. The application of AI in education can be traced back to the 1950s when the first generation of Intelligent Tutoring Systems (ITS) emerged, which were rule-based expert systems used to simulate the knowledge and strategies of human teachers. With the development of computer technology and AI theories, the application of AI in education has continuously innovated and expanded, covering intelligent textbooks, intelligent assessments, intelligent tutoring, and intelligent experiments.

AI can provide effective technical support for STEM education, including:

1. Intelligent Textbooks: AI can dynamically generate or adjust STEM textbook content, difficulty, and formats based on students' levels, interests, and styles to meet individual needs.

2. Intelligent Assessment: AI can use natural language processing, image recognition, and data mining technologies to automatically or semi-automatically grade or provide feedback on students' assignments, projects, and experiments in STEM education, improving assessment efficiency and quality.

3. Intelligent Tutoring: AI can build models of students and domains using machine learning and knowledge representation technologies, providing personalized guidance or interventions to promote cognitive and emotional development in STEM education.

4. Intelligent Experiments: AI can use virtual reality (VR) and augmented reality (AR) technologies to provide realistic STEM experiment scenarios and operations, enhancing students' practical abilities and exploratory spirit.

In summary, the application of AI in STEM education is diverse and innovative, offering effective technical support and solutions. However, it also faces technical, ethical, and social challenges that require further research and discussion.

#### AI-Driven Personalized Learning

Personalized learning refers to educational methods that provide suitable learning content, pathways, strategies, and feedback for each student based on their individual differences and learning needs to improve learning outcomes and satisfaction. AI is a crucial technological means for achieving personalized learning, as it can collect and analyze vast amounts of learning data to build and update models of students and domains, thereby providing customized learning plans and support for each student. Specifically, AI can promote personalized learning in the following ways:

1. Personalized Learning Content: AI can dynamically generate or recommend suitable learning content, including textbooks, exercises, cases, and videos, based on students' prior knowledge, skill levels, and interests.

2. Personalized Learning Pathways: AI can dynamically plan or adjust each student's learning pathway, including the sequence, difficulty, and duration of learning activities, based on their goals, styles, and pace.

3. Personalized Learning Strategies: AI can provide or suggest personalized learning strategies, including prompts, encouragement, and incentives, based on students' cognitive, emotional, and social factors.

4. Personalized Learning Feedback: AI can dynamically provide or adjust learning feedback, including evaluations, suggestions, and rewards, based on students' performance, needs, and expectations.

AI promotes personalized learning in various ways, enabling each student to receive a tailored learning experience and support. Personalized learning is significant for STEM education as it can enhance students' academic performance and interest, foster creativity and critical thinking, and meet diverse learning needs.

#### AI Empowering STEM Education

As AI deeply integrates with STEM education, it continuously refines STEM educational concepts. Educators can use various digital tools and channels to access systematic, scientific, and up-to-date STEM educational views. They will further recognize that STEM education is for all learners, based on project and problem contexts, advocating "learning by doing," relying on tools and resources, and involving multiple stakeholders in a new educational model. The conceptual system of STEM education will inevitably become more mature, open, and diverse with the rapid development of AI.

AI will also enable iterative updates of STEM education components. STEM education mainly cultivates and develops learners' interdisciplinary knowledge, problem discovery and solving abilities, collaborative inquiry skills, and practical innovation capabilities. From a longitudinal perspective, AI sets higher requirements for the STEM literacy of learners and educators. Students must possess richer interdisciplinary knowledge, stronger independent thinking and collaborative inquiry abilities, and a keener awareness of problem discovery to cope with increasingly complex and variable problem situations. From a horizontal perspective, AI also demands that students have the necessary qualities for learning in an intelligent environment, including awareness, knowledge, abilities, technologies, methods, and thinking related to intelligent learning.

In an AI-empowered STEM education context, learners can receive help from intelligent tutors and online experts, making remote autonomous learning and collaborative inquiry easier.

It is worth noting that STEM education is not a simple addition of science, technology, engineering, and mathematics education, but an organic integration of these disciplines. The traditional "siloed" subject teaching with only a few cross-disciplinary integrated practice courses will gradually give way to exclusive STEM courses. Virtual reality or augmented reality technology's scenario creation, personalized guidance and assistance from intelligent tutors, and remote collaboration from external forces will also provide more curriculum resources and model choices for STEM teaching.

#### Theoretical Framework

This research explores the application and impact of AI in personalized STEM education from the perspective of intelligent education. To establish a comprehensive theoretical framework, this study integrates intelligent education theory, personalized learning theory, and STEM education theory.

1. Intelligent Education Theory: Intelligent education is an innovative education model that uses next-generation information technology to build an ecological learning environment and cultivate creative talents. It includes the connotation and characteristics, principles, and methods of intelligent education.

2. Personalized Learning Theory: Personalized learning is an educational model that provides resources, pathways, and support tailored to each learner's characteristics, needs, and goals, allowing them to learn effectively at their own pace and achieve their maximum potential. This theory covers the connotation and characteristics, theoretical foundation, and implementation strategies of personalized learning.

3. STEM Education Theory: STEM education aims to cultivate innovative and practical talents. It emphasizes cross-disciplinary integration and application, as well as problem-based inquiry and solutions. This theory includes the connotation and characteristics, theoretical foundation, and implementation strategies of STEM education.

4. Comprehensive Theoretical Framework: Combining intelligent education theory, personalized learning theory, and STEM education theory, a comprehensive theoretical framework is constructed, including research background, research objectives, theoretical foundation, and research content.

#### Application Research

Based on the theoretical framework mentioned above, this research analyzes four types of AI applications in personalized STEM education: AI as a STEM content generator, evaluator, tutor, and experimenter, and evaluates their advantages and disadvantages.

1. AI as a STEM Content Generator: AI generates or adjusts textbook content, difficulty, and format to meet students' individual needs. The advantage is improving the quality and diversity of textbooks to meet different learning differences, while the disadvantage is that it may contain errors or inappropriate content, requiring teacher review and modification.

2. AI as a STEM Evaluator: AI conducts assessments of learning outcomes or processes to improve assessment efficiency and quality. The advantage is enhancing evaluation efficiency and effect, reducing teacher workload, while the disadvantage is that assessment standards or methods may not fit well and need adjustment and supplementation by teachers.

3. AI as a STEM Tutor: AI provides learning guidance or support to promote cognitive and emotional development. The advantage is improving the availability and flexibility of tutoring, boosting students' learning confidence and ability, while the disadvantage is that the content or method of tutoring may be inaccurate or inappropriate, requiring teacher supervision and intervention.

4. AI as a STEM Experimenter: AI provides experimental scenarios and operations to enhance students' practical abilities and exploratory spirit. The advantage is improving the realism and fun of experiments, stimulating students' interest and creativity, while the disadvantage is that experimental content or methods may be incomplete or unsafe, needing teacher guidance and monitoring.

AI's broad and significant applications in personalized STEM education, from content generation to evaluation, tutoring, and experimentation, can promote personalized learning in various ways. However, these applications also have limitations and risks that teachers and students need to assess and balance carefully.

#### Impact of AI on Personalized STEM Education

1. Positive Impact:

- Efficiency: AI optimizes teaching and learning tasks and processes through automation, enhancing the efficiency

of personalized STEM education.

- Effectiveness: AI provides personalized and adaptive teaching and learning support, improving the effectiveness of personalized STEM education.

- Accessibility: AI offers open and shared teaching and learning platforms and resources, increasing the accessibility of personalized STEM education.

- Flexibility: AI provides diverse and flexible teaching and learning methods, enhancing the flexibility of personalized STEM education.

- Realism: AI offers simulated and virtual teaching and learning environments, improving the realism of personalized STEM education.

- Engagement: AI provides gamified and entertaining teaching and learning activities, enhancing the engagement of personalized STEM education.

- Adaptability: AI provides dynamic and intelligent teaching and learning adjustments, enhancing the adaptability of personalized STEM education.

- Innovation: AI offers autonomous and innovative teaching and learning opportunities, enhancing the innovation of personalized STEM education.

2. Potential Negative Impact:

- Data Privacy: AI's collection and analysis of student learning data may involve privacy and security issues.

- Algorithm Bias: AI algorithms may lack transparency or explainability, affecting the accuracy and reliability of the system.

- Human-AI Collaboration: Effective and balanced human-AI collaboration may be challenging, affecting the willingness and ability of teachers and students to collaborate.

- STEM Education Integrity: AI's compatibility with the essence and goals of STEM education may face inconsistencies or mismatches.

3. Balancing the Positive and Negative Impacts of AI in Personalized STEM Education:

- Establishing Standards: Developing standards to evaluate AI's impact on personalized STEM education.

- Enhancing Literacy: Improving the AI literacy and capabilities of teachers and students.

- Strengthening Supervision: Enhancing supervision and feedback on processes and outcomes.

The above measures and strategies can help balance the positive and negative impacts of AI in personalized STEM education and promote the development of future personalized learning.

#### Conclusion and Future Research Directions

1. Key Findings and Conclusions:

This study delves into the application and impact of AI in personalized STEM education. The background and significance of this research are self-evident. In modern education, the use of AI has become a crucial topic, especially in STEM education. The main findings of this study are:

- Widespread Application of AI in STEM Education: From literature reviews, this study found that numerous AI tools and technologies are applied in STEM education, including adaptive learning systems, intelligent tutoring systems, and virtual laboratories. These tools and technologies not only enrich teaching methods in STEM education but also effectively improve students' learning efficiency and outcomes.

- AI Drives the Development of Personalized Learning: AI technology can collect and analyze learning data to provide targeted learning resources and strategies for each student, thus achieving personalized learning.

- AI Has Positive Effects on Personalized STEM Education but Also Potential Negative Impacts: On the one hand, AI can enhance teaching efficiency and provide personalized learning resources to increase students' motivation. On the other hand, AI use may lead to privacy breaches and over-reliance on technology, among other issues.

Therefore, balancing the positive and negative impacts of AI in personalized STEM education is a topic worthy of further research. In balancing these impacts, we propose a series of recommendations and strategies, including establishing standards, improving literacy, and strengthening supervision.

2. Future Research Directions:

As AI technology continues to develop and improve, its application and impact on personalized STEM education will expand and deepen. Therefore, future research directions should consider the following aspects:

- In-Depth Research on AI's Role in More STEM Education Fields: Future studies could explore how AI can play a role in various levels (e.g., primary, secondary, and higher education) and fields (e.g., science education, technology education, engineering education, and mathematics education) of STEM education.

- More Effective Use of AI Technology for Personalized Teaching: Future research can investigate how to design and develop more efficient, accurate, fair, safe, and adaptive AI systems to better support and promote personalized STEM education.

- Addressing Ethical and Moral Issues in AI Technology Application: AI might lead to issues like data privacy breaches, algorithm discrimination, and technological dependence. Future research should explore how to resolve these issues comprehensively.

??????????????-?????????????????? ????????????????:

???? ???? ?????????????????? ???? ?????????????????? ???????? ???????? ??????????????????, ?????? ?????? ???? ?????????????? ?????? ???????????????? ???? ???????????????????????? ???????????????? ?????? ???????????????????? ???????? ?????? ?????????????? ???????????????????? ???? ???????? ??????????????, ?????????????????? ????????, ?????? ?????? ?????????????????? ????????-???????????????? ???? ???????????????????? ???? ?????? ?????????????????????