Human-Centred Learning Analytics and AI in Education: a Systematic Literature Review

Intelligent Tutoring Systems and Automated Feedback

Context: Educational settings including higher education, K-12, and medical education, targeting students and teachers in various subjects such as computer science, writing, and surgical training.

Implementation: Integration of intelligent tutoring systems and automated feedback mechanisms using AI technologies to provide personalized learning support, timely feedback, and data-driven insights through interactive platforms like dashboards and tutoring systems.

Outcomes: Enhanced personalized learning experiences, improved tracking of student progress, increased writing proficiency, and better decision-making skills, leading to improved educational outcomes and student engagement.

Challenges: Concerns about data privacy, algorithmic bias, resistance from educators, the accuracy of feedback, integration with existing curricula, and the complexity of specific domains such as surgical decision-making.

Collaborative Chatbots for Interactive Learning

Context: Contexts include parent-child interactive storytelling, as well as various educational environments aimed at enhancing engagement through AI interactions.

Implementation: Development of chatbots that facilitate collaborative storytelling and feedback interactions between students, parents, and educators, utilizing AI technologies to enhance user engagement.

Outcomes: Improved parental involvement and student engagement in educational activities, leading to a more interactive learning experience.

Challenges: Balancing AI involvement with human input, ensuring a seamless user experience, and addressing diverse student needs.

Personalized Course Recommendation Systems

Context: Higher education, assisting students with course selection based on their preferences and interests.

Implementation: Development of recommendation systems that analyze student preferences and provide tailored course suggestions to improve alignment with student interests.

Outcomes: Increased student satisfaction and improved course alignment with students' academic and career goals.

Challenges: Data privacy concerns and the need for system adaptability to cater to diverse student populations.

Teacher Dashboards for Real-Time Instructional Decisions

Context: Middle school STEM curriculum and K-12 educational settings, where teachers need to customize lesson plans based on student data.

Implementation: Use of human-centered learning analytics to develop scalable dashboards that provide teachers with insights into student performance and engagement.

Outcomes: Enhanced lesson personalization, improved teacher decision-making, and better student learning experiences through data-driven insights.

Challenges: Ensuring data privacy, usability for teachers, and integration with existing curricula.

Technical Barrier

Lack of involvement from educational stakeholders in the design and implementation phases of generative AI systems, as well as challenges in maintaining a balance between human control and computer automation, and integration challenges with existing educational systems and curricula.

Proposed Solutions: Implement participatory design methods to engage students and teachers actively throughout the design process. Utilize frameworks like the HCAI framework to assess and balance features of control and automation. Develop modular AI systems that can be easily integrated into current platforms. Investing in professional development for educators on AI tools and fostering collaboration between technologists and educators.

Ethical Barrier

Concerns regarding data privacy and the potential for algorithms to perpetuate biases and discrimination, as well as ethical use of AI in educational settings.

Proposed Solutions: Adopt transparent data collection practices and ensure ethical guidelines in AI system design. Establish clear privacy policies and ethical guidelines for AI use in education, incorporating transparency in AI systems.

Cultural Barrier

Resistance to adopting AI tools among educators who prefer traditional teaching methods.

Proposed Solutions: Provide professional development and training to demonstrate the benefits of AI in enhancing teaching.

Usability Barrier

Challenges in ensuring that AI tools are user-friendly for educators and students.

Proposed Solutions: Conduct co-design workshops with end-users to ensure tools are tailored to their needs.

Project Contact: Dr. Jianhua Yang

LLM Model Version: gpt-4o-mini-2024-07-18

Analysis Provider: Openai