New Applications of AI in STEM Education
Introduction
Artificial intelligence (AI) has become increasingly influential in human lives today, shaping industries, enhancing convenience, and transforming how society works, communicates, and makes decisions. This new technology has also been the source of many ethical discussions, including bias and accessibility.
Through the evolution of AI, educators and programmers have been working together to create systems that will enhance student education and ability to learn in new, revolutionary ways. AI in education, specifically science, technology, engineering, and math (STEM), has enhanced personalized learning, streamlined complex problem-solving in subjects such as chemistry and physics, and provided students with interactive and adaptive tools that cater to their unique learning styles.
In 2022, the International Journal of STEM Education published an article that compiled 63 studies surrounding AI in STEM education. These studies discussed the potential future applications of AI in education and its benefits and drawbacks (Xu & Ouyang, 2022). Studying a range from kindergarteners to undergraduate students, five AI innovations were found in the field of STEM education as potentially beneficial to the current education system.
Learning Prediction
Making up 29% of new learning innovations, Learning Prediction is a method used to make calculated projections of students’ academic trends, allowing AI to give them the best possible suggestions to maximize learning (Xu & Ouyang, 2022). Learning Prediction works by giving an AI engine aspects of a student’s attributes, such as level of attention in class, performance, and class behaviors, to accurately predict a student’s performance. This can positively impact education by giving the teacher more insights as to what teaching techniques are personalized, and most effective to which students, such as traditional lectures, problem-based learning, project-based learning, game-based learning, self-learning, and collaborative learning.
One example is gathering homework grades, grade point averages, and study methods to predict students’ future assessments. According to a 2019 study, Cabot Zabriskie and his team used Learning Prediction to help college students taking Physics one and two classes and discovered that after a week of using the AI learning tool, there was an 80% accuracy rate, a 12% accuracy increase (Zabriskie et al., 2019). This means that AI can benefit STEM education by revealing which practices are most effective in getting higher grades among students. Learning Prediction’s software is also important for teachers to track their students’ progress and give more specialized attention to those who want to improve their grades.
Intelligent Tutoring Systems (ITS)
The second innovation studied in Xu & Ouyang’s article is Intelligent Tutoring Systems (ITS). Xu & Ouyang’s innovation learns from students’ input, such as their preferred study method, their intended pace, and goals, and gives them recommendations for ways to study that will best set them up for success (Xu & Ouyang, 2022). This is an independent system where the student employs the AI tutor. ITS programs provide feedback on what study resources and techniques students should use in the classroom, such as one-on-one tutoring and interactive AI systems that adapt to a class’s pace and level.
AI has the ability to create lesson plans and activities that are tailored to a student’s unique learning style to help them learn best. For example, an adaptive AI model was used to teach students about the physics surrounding pulleys (Myneni et al., 2013). These simulations allow for problem-solving and better visualization of concepts that are difficult to understand in real life, as well as adapting to a student’s knowledge and learning skills. This fosters a more engaging environment where students learn at a pace that is best for them.
Student Behavior Detection
Student Behavior Detection focuses on identifying certain student behaviors that inform teachers the best way to teach each student most effectively. For example, using a system that tracks a student’s effort and desire to review content through teacher input, diligence, and reviewing patterns to understand how they work and the future implications of their studying (Hsiao et al., 2020). This process is similar to learning prediction but has a greater focus on observing and understanding students’ behaviors rather than test scores.
Automation
Historically, teachers have had to spend a lot of time working on creating tests as well as grading them. AI has improved classroom dynamics by taking academic information from teachers, such as progress reports and test scores, to streamline skills evaluations (Aldabe & Maritxalar, 2014). It is also helpful for teachers to learn new questions to ask for assessments and allow students to anticipate questions that could appear on assessments. This helps teachers ease their responsibilities and allows students more variety when studying for assessments. Automation makes AI valuable in education today because it improves the quality of work for both teachers and students, allowing for a better standard of education in STEM.
Educational robots
Hands-on activities have been very useful to engage students and foster growth in classes that focus on technology and engineering by using innovative strategies. Programming tools such as LEGO MINDSTORMS EV3 help enrich a student’s problem-solving skills through the use of social robots that mimic human behavior while keeping a focus on cooperation and learning. This is beneficial to students because it keeps them focused on the task at hand and learn at the best pace for them (Belpaeme et al., 2018).
Additional Innovations
Additional innovations were found in Xu & Ouyang’s study, such as an AI model that created student study groups in a college class based on individual characteristics that the professor input into the AI model (Tehlan et al., 2020). This increased cooperation among students because their compatible personalities made the learning process quicker and more effective.
Education through AI is also proven to be effective in educating students with learning disabilities (Kohnke & Zaugg, 2025). AI models are able to foster learning with students with learning disabilities, specifically students with Autism Spectrum Disorder (ASD) and Attention-Deficit/Hyperactivity Disorder (ADHD) by empowering student engagement and supporting critical thinking and problem solving, while also addressing barriers students might have because of AI’s capacity to tailor one’s education.
Ethical Concerns
While AI education in STEM has a bright future, there are still many concerns about its implementation. One concern is the reliability of AI's functions and whether the information it gives students and teachers is credible, since AI could undergo unauthorized personal data collection. Ethical concerns such as bias and privacy risks are topics to focus on and are cause for debate and consideration for future implications (Kohnke & Zaugg, 2025). Another ethical issue is that access to AI tutoring and other innovations is not universally available, prompting discussions on their use. This is because some say that it is not fair that certain students get access to these tools while others cannot because of their school districts or financial situation.
Conclusion
While there is still a lot of learning and innovation to come within AI, there is a large upside and a bright future in implementing AI learning into today’s education. Students have been showing positive results from using AI in their learning, creating an environment that will welcome the future of innovation into the world of education.
References
Joseph, N. O. B., & Uzondu, N. N. C. (2024). Integrating AI and Machine Learning in
STEM education: Challenges and opportunities. Computer Science & IT Research
Journal, 5(8), 1732–1750. https://doi.org/10.51594/csitrj.v5i8.1379
Kohnke, S., & Zaugg, T. (2025). Artificial intelligence: an untapped opportunity for equity and access in STEM education. Education Sciences, 15(1), 68.
https://doi.org/10.3390/educsci15010068
Myneni Lakshman. An interactive and intelligent learning system for physics education. (n.d.). IEEE Journals & Magazine | IEEE Xplore. https://ieeexplore.ieee.org/document/6559985
Smith-Mutegi, D., Mamo, Y., Kim, J., Crompton, H., & McConnell, M. (2025). Perceptions of STEM education and artificial intelligence: a Twitter (X) sentiment analysis. International
Journal of STEM Education, 12(1). https://doi.org/10.1186/s40594-025-00527-5
Xu, W., & Ouyang, F. (2022). The application of AI technologies in STEM education: a
systematic review from 2011 to 2021. International Journal of STEM Education, 9(1).
https://doi.org/10.1186/s40594-022-00377-5
Zabriskie, C., Yang, J., DeVore, S., & Stewart, J. (2019). Using machine learning to predict physics course outcomes. Physical Review Physics Education Research, 15(2).