Artificial Intelligence Empowers Primary School Students’ STEM Education: A Review of Independent Learning Interests in Hong Kong, 2024–2025
DOI:
https://doi.org/10.61173/4yvc1843Keywords:
Artificial Intelligence, Primary School Students, STEM EducationAbstract
Educational equity is an emphasis of the United Nations. Usually, due to the economic backwardness in remote areas, serious educational inequality problems still exist, especially in some developing countries, including China. Through accompanying learning and answering questions, it helps alleviate the problem of insufficient educational resources for students in remote areas and further enhances students’ autonomous learning ability. However, even educational. Therefore, in this study, the systematic review method was adopted. Based on the summary of 12 core literature screened out, it was found that the primary stem research was conducted in Hong Kong. The results show that it can help enhance students’ autonomous learning ability and provide timely personal feedback, thereby improving students’ autonomous learning ability. However, in the future, it is necessary to set a reasonable scope for their use during their learning process. This study is the latest summary of research on the management of software use in the field of education in Hong Kong.
References
Overall, this study used a systematic review approach to Unesco.org, 2024. https://unesdoc.unesco.org/ark:/48223/ collect and analyze research from the past three years on pf0000391104 the progress and challenges of AI applications in STEM [2] UNESCO, “Artificial intelligence in education,” UNESCO,
education in primary schools in Hong Kong. Four key 2023. https://www.unesco.org/en/digital-education/artificialfindings emerged: First, to date, research on the applica- intelligence tion of LLM-based AI tools in primary education has pri- [3] X. Zhai et al., “A Review of Artificial Intelligence
marily focused on machine learning platforms/software, (AI) in Education from 2010 to 2020,” Complexity, vol.
particularly ChatGPT, as well as other tools like Claude 2021, no. 8812542, pp. 1–18, Apr. 2021, doi: https://doi. and Bard. These tools typically use natural language pro- org/10.1155/2021/8812542. cessing and generation to provide personalized learning [4] W. Xu and F. Ouyang, “The Application of AI Technologies
support for students, including language learning and in STEM education: a Systematic Review from 2011 to 2021,” mathematics. Second, all ten core studies found that LLM International Journal of STEM Education, vol. 9, no. 1, pp. 1–20,
AI tools like ChatGPT can enhance student learning inter- Sep. 2022, doi: https://doi.org/10.1186/s40594-022-00377-5. est by providing timely answers to student questions and [5] W. J. Triplett, “Artificial Intelligence in STEM Education,” simplifying complex knowledge points. Furthermore, they Cybersecurity and Innovative Technology Journal, vol. 1, no. 1,
can provide personalized suggestions based on students‘ pp. 23–29, Sep. 2023, doi: https://doi.org/10.53889/citj.v1i1.296. learning abilities and daily questions, further enhancing [6] M. Ali, “State of STEM Education in Hong Kong: A
student autonomy in learning. Policy Review,” Academia Letters, Oct. 2021, doi: https://doi. Thirdly, in the research on primary school students in org/10.20935/al3680. Hong Kong so far, although it has been confirmed that [7] Y. Song, “Redefining STEM Education in the Post-ChatGPT tools can promote the improvement of autonomous learn- Era—Case Studies and Perspectives,” SSRN Electronic Journal,
ing ability, it is necessary to do so under the premise of Jan. 2024, doi: https://doi.org/10.2139/ssrn.4733685. providing correct strategic guidance, especially in the [8] T. Wang and E. C. Keung Cheng, “An investigation of formulation of learning plans, monitoring, reflection, and barriers to Hong Kong K-12 schools incorporating Artificial correction, and cooperative learning. education students in Intelligence in education,” Computers and Education: Artificial
Hong Kong primary schools, particularly in terms of stu- Intelligence, vol. 2, p. 100031, Aug. 2021, doi: https://doi. dents‘ academic performance, critical thinking, and long- org/10.1016/j.caeai.2021.100031. term outcomes. Specifically, in terms of academic perfor- [9] E. Chng, A. L. Tan, and S. C. Tan, “Examining the Use mance, through personalized guidance, students‘ cognitive of Emerging Technologies in Schools: a Review of Artificial load is effectively reduced, and their learning outcomes Intelligence and Immersive Technologies in STEM Education,”
and the speed of knowledge internalization are enhanced. Journal for STEM Education Research, vol. 6, Apr. 2023, doi: In terms of critical thinking, it is found that students can https://doi.org/10.1007/s41979-023-00092-y. make use of the diverse learning materials provided by [10] Y. Dai, Q. Panghe, Y. Zhang, M. Zhang, and X. Xu, “How AI‘s real-time feedback, which is helpful for them to LLMs Support EFL Writing:A Case Study of K-12 English conduct some critical thinking management training and Learning Based on the EDIPT Model,” 2024 International promote the development of deep learning and innovation Conference on Intelligent Education and Intelligent Research
capabilities. In terms of long-term effectiveness, in the (IEIR), pp. 1–8, Nov. 2024, doi: https://doi.org/10.1109/ future, students are encouraged to use it for autonomous ieir62538.2024.10959858. learning. However, it is necessary to regularly update the [11] E.Dimitriadou and A. Lanitis, “A critical evaluation, accuracy and information illusion issues of the tools them- challenges, and future perspectives of using artificial intelligence selves, and call on students to avoid excessive reliance on and emerging technologies in smart classrooms,” Smart Dean&Francis Rongyu Yang
Learning Environments, vol. 10, Art. no. 12, 2023, doi: https:// doi.org/10.1145/3613904.3642198. doi.org/10.1186/s40561-023-00231-3 [15] M. Rahman et al., “ChatGPT in Research and Education: A [12] U. Lee et al., “I see you: teacher analytics with GPT- SWOT Analysis of Its Academic Impact,” Computer Modeling 4 vision-powered observational assessment,” Smart Learning in Engineering & Sciences, vol. 0, no. 0, pp. 1–10, Jan. 2025,
Environments, vol. 11, no. 1, Oct. 2024, doi: https://doi. doi: https://doi.org/10.32604/cmes.2025.064168. org/10.1186/s40561-024-00335-4. [16] H. Li, R. Xiao, H. Nieu, Y.-J. Tseng, and G. Liao, “‘From [13] X. Wei, L. Wang, L.-K. Lee, and R. Liu, “Multiple Unseen Needs to Classroom Solutions’: Exploring AI Literacy Generative AI Pedagogical Agents in Augmented Reality Challenges & Opportunities with Project-Based Learning Toolkit Environments: A Study on Implementing the 5E Model in in K-12 Education,” Proceedings of the AAAI Conference on Science Education,” Journal of Educational Computing Artificial Intelligence, vol. 39, no. 28, pp. 29145–29152, Apr.
Research, vol. 63, no. 2, Dec. 2024, doi: https://doi. 2025, doi: https://doi.org/10.1609/aaai.v39i28.35187. org/10.1177/07356331241305519. [17] Q. Zhu, M. Wang, T. Zhang, and H. Huang, “Current Trends [14] Dylan Edward Moore, S. Moore, B. Ireen, W. P. and Future Prospects of Large-Scale Foundation Model in K-12 Iskandar, G. Artazyan, and E. L. Murnane, “Teaching artificial Education,” Frontiers of digital education., vol. 2, no. 2, May
intelligence in extracurricular contexts through narrative-based 2025, doi: https://doi.org/10.1007/s44366-025-0059-6.
learnersourcing,” ACM Digital Library, May 2024, doi: https://
Downloads
Published
Issue
Section
License
Copyright (c) 2025 by the authors.
This work is licensed under a Creative Commons Attribution 4.0 International License.
