Artificial intelligence in the classroom
How AI can support teachers and students
Many teachers and educational leaders use artificial intelligence (AI) daily, from getting music and film recommendations on streaming services to using smart speakers.
Yet, few use AI in the classroom beyond the built-in AI features in learning and school management systems.
Challenges
Some educators' skepticism of AI may stem from the launch of the chatbot ChatGPT in late 2022.
This made generative AI (applications that can create text, audio, and video) accessible to a wide audience.
This caused concern among schools, universities, and policymakers. Would students use tools like ChatGPT to cheat in exams and assignments?
Challenges such as cost, privacy concerns, security risks, and the potential impact on students' creativity and critical thinking skills may also be factors in the slow takeup of AI in teaching.
Embracing AI in schools
Despite possible downsides, teachers and educational leaders are embracing AI in the classroom while managing its risks.
In this article, we're highlighting five areas where AI can support teachers and empower students – now and as the technology develops:
- Supporting students with disabilities
- Aiding independent study
- Lesson planning
- Classroom management
- Grading and feedback
We also highlight how schools and policymakers can minimize AI's risks to teaching and learning.
1. Supporting disabled students
Teaching a group of children with differing learning styles is challenging. It's even more difficult when the group includes students with physical or learning disabilities.
This review by Dr. Mary F. Rice and Dr. Shernette Dunn in the journal Computers in the Schools highlights how AI in assistive technology can support students with disabilities including autism, dyslexia, Down Syndrome, visual disabilities, and physical disabilities.
The review mentions applications that can help identify learners who struggle with reading and help autistic children improve their social skills.
Low cost
Schools don't necessarily need access to expensive or specialist tools to use AI to support students.
In the book Artificial Intelligence in Schools Varun Arora highlights free and inexpensive tools such as:
- Google Read Along, which provides feedback on children's reading
- Microsoft's Seeing AI, which uses a mobile device's camera to describe the user's environment including people, colors, and objects
Text-to-speech apps, which are freely available on many devices, can be useful tools for students who struggle with speech or find it difficult to vocalize their thoughts.
2. Aiding independent study
In AI for Learning, the authors remind us that AI-based teaching and tutor systems have been available since the 1970s. Nowadays, online generative AI tools like ChatGPT and Microsoft Copilot can also help children study independently.
For example, "students can also use [chatbot] queries to have themes and key concepts explained to them in a way that is much more dynamic and personalized than a traditional textbook glossary or dictionary."
AI chatbots can be especially useful for teaching languages. This study found that ChatGPT can be a powerful tool for teaching students the intricacies of English grammar and sentence construction.
It would not be wrong to place AI in the camp with calculators, computers, and Wikipedia as just another tool for learning
Downsides
The study also highlights that students and teachers cannot completely trust the accuracy of ChatGPT's content. As well as this, teachers in the study expressed concerns that using generative AI could prevent learners from developing writing and critical thinking skills.
A similar study concluded that using generative AI tools as a study aid requires careful task design to make sure students stay motivated and interested in their learning.
Meanwhile, the authors of AI for Learning say using AI as a teaching assistant should be a "'human-AI transactive collaboration, each with their own clear roles."
3. Lesson planning
"To my surprise, [ChatGPT] gave me an excellent lesson guide with activities, time allotment, materials, and lesson procedures."
That's the verdict from a teacher who was part of a study on teachers' experiences of using ChatGPT in learning. The study highlights how teachers are using generative AI tools to support them in creating lesson plans that include activities, resources, and instructions.
Another study, which examined the role of AI in special education, found that generative AI can be useful for supporting teachers in creating customized plans for students with specific needs.
An article for the Journal of Political Science Education by Dr. Steven Michels highlights how teachers can use generative AI to produce first drafts of teaching sessions or make lesson content more inclusive.
Limitations
Importantly, these studies also highlight the limitations of using these tools, notably the potential biases and inaccuracies in the content they generate.
4. Classroom management
Managing a class of children is often complex. Teachers must not only engage students with learning materials, but also manage behavior and performance, build relationships, and create an inclusive learning environment.
In a chapter in Artificial Intelligence in STEM Education, researchers call this "classroom orchestration." They highlight how AI tools can support teachers' classroom orchestration to help them:
- Use their cognitive capacity and expertise more effectively
- Direct their attention to students who are most in need
The researchers then describe an AI tool that supports teachers by:
- Monitoring and analyzing student progress in real time
- Automating administrative tasks that don't need the teacher's input
Using AI to help manage student behavior in the classroom is the subject of a chapter in Varun Arora's book Artificial Intelligence in Schools.
He highlights how AI is getting better at:
- Identifying different speakers in a large group, even when people talk over each other
- Understanding emotion and sentiment in speech
- Understanding meaning and emotion in physical actions
This means that AI could monitor a whole classroom, and alert busy teachers to potential behavior issues early on, based on what students are saying and doing. Teachers can then decide what interventions – if any – are needed.
However, this raises ethical and privacy concerns that schools must manage before they implement this type of technology.
5. Grading and feedback
Teachers spend a lot of time reviewing and grading their students' work. Much of this happens away from the classroom, often during what could be teachers' personal time.
In recent years, some schools have provided teachers with AI tools that support them with grading and marking.
These tools are currently most effective on "lower order" questions, which encourage students to memorize information rather than do analytical thinking.
Personalizing feedback
Teachers can also use AI tools to provide personalized, constructive feedback.
In this study, researchers created a web application, integrated with ChatGPT. They instructed ChatGPT, via a prompt, to provide feedback that aligns with established feedback principles.
Students put their answers to questions into the application and received feedback via ChatGPT.
The researchers concluded that AI tools such as ChatGPT "can significantly enhance the scalability and consistency of feedback."
However, they also found that the tool occasionally "missed opportunities to identify misconceptions or provide comprehensive, actionable advice."
Further risks
Researchers have highlighted other risks of relying on AI to grade student work.
These include worsening inequality in education, and concerns about AI replacing human educators.
Realizing the benefits of AI in the classroom
To realize the benefits of AI in the classroom while minimizing its risks, schools and policymakers can implement safeguards and policies that:
- Identify and address ethical and privacy concerns
- Encourage critical thinking, problem-solving, and creativity in the classroom
- Ensure new teaching practices introduced by AI tools are based on evidence
- Maintain a balance between technological and social interactions
Further reading:
Books
AI for Learning
By Carmel Kent and Benedict du Boulay
AI for School Teachers
By Rose Luckin, Karine George, and Mutlu Cukurova
Artificial Intelligence Applications in K-12: Theories, Ethics, and Case Studies for Schools
Edited by Helen Crompton and Diane Burke
Artificial Intelligence in STEM Education: The Paradigmatic Shifts in Research, Education, and Technology
Edited by Fan Ouyang, Pengcheng Jiao, Bruce M. McLaren, and Amir Alavi
School Children and the Challenge of Managing AI Technologies: Fostering a Critical Relationship Through Aesthetic Experiences
Edited by Emanuela Guarcello and Abele Longo
The Ethics of Artificial Intelligence in Education: Practices, Challenges, and Debates
Edited by Wayne Holmes and Kaśka Porayska-Pomsta
Journal articles
- Artificial intelligence and K-12 education: possibilities, pedagogies and risks by Joseph Mintz, Wayne Holmes, Leping Liu and Maria Perez-Ortiz in Computers in the Schools
- Artificial intelligence pedagogical chatbots as L2 conversational agents by Assim S. Alrajhi in Cogent Education (open access)
- Exploring the use of artificial intelligence (AI) in the delivery of effective feedback by Julia Venter, Stephen A. Coetzee and Astrid Schmulian in Assessment & Evaluation in Higher Education (open access)
- Teaching (with) artificial intelligence: the next twenty years by Steven Michels in Journal of Political Science Education
- The use of artificial intelligence with students with identified disabilities: a systematic review with critique by Mary F. Rice and Shernette Dunn in Computers in the Schools
- To generate or stop generating response: exploring EFL teachers' perspectives on ChatGPT in English language teaching in Thailand by Mark Bedoya Ulla, William F. Perales and Stephenie Ong Busbus in Learning: Research and Practice
- Transforming educational assessment: insights into the use of ChatGPT and large language models in grading by Chokri Kooli and Nadia Yusuf in International Journal of Human–Computer Interaction