Using AI to foster self-directed learning

Learning challenges often arise from a lack of self-direction, engagement and motivation. To address these, we integrated an AI-powered learning platform into our Moodle-based course. The centrepiece of this platform is a sophisticated chatbot named Smart Agent, designed to foster self-directed learning and offer personalised support in listening, reading, writing and speaking.

Strategies for implementation:

1. Creating short-term goals

We restructured our curriculum by breaking down learning outcomes and tasks into achievable short-term goals. We displayed these prominently on our online learning platform to provide students with a clear road map for their learning journey.

For example, the learning outcome of one listening task was to summarise a short video. To support students, we divided the task into seven incremental steps:

• Listen for missing words
• Outline the structure
• Take notes
• Write a short summary
• Ask Smart Agent to summarise the video content using the transcript
• Ask Smart Agent to compare both summaries
• Ask Smart Agent for suggestions on what to focus on based on each

We provided specific guidance to students who missed words in the listening task, directing them to relevant parts of the video (e.g., "You missed the fourth word; please watch the video from 2:14–2:29 again for the answer"). This targeted approach helps students focus on areas needing improvement.

Additionally, we developed structured templates to guide students in writing summaries. These templates helped students organise their thoughts and include all necessary elements.

Since students often struggle to phrase effective AI queries, we created a question bank containing well-crafted questions that students can use or adapt when engaging with the Smart Agent chatbot. This approach helps students learn how to ask questions that elicit useful responses from AI, enhancing their ability to extract relevant information and guidance.

To keep students motivated, we implemented a visual progress tracker on the learning platform, which included scores for each step, checklists for tasks and badges for completed stages. This helped students monitor their progress.

• Spotlight guide: Bringing GenAI into the university classroom
• An insider’s guide to how students use GenAI tools
• Can AI offer everyone a personal tutor 24/7?

2. Embedding an AI Chatbot into the learning platform

We integrated Smart Agent into our Moodle platform through the following steps:
a) Selecting an appropriate AI chatbot framework: We researched and selected a chatbot solution compatible with Moodle, such as Dialogflow by Google or MoodleBot, which are specifically designed for Moodle integration.
b) Creating a comprehensive dataset: We compiled a database of course materials, including lecture transcripts, assignment instructions and key concepts to train the chatbot. We also developed a list of frequently asked questions and appropriate responses to ensure the chatbot could provide relevant information.
c) Rigorous testing: We conducted extensive testing to ensure the chatbot functioned effectively and provided a positive user experience.

Tips for implementation:
a) Start with a small-scale pilot to identify and resolve technical issues before full implementation.
b) Regularly update the chatbot’s dataset based on student interactions and feedback to improve its accuracy and relevance.

3. Implementing biweekly reflection sessions

To complement the AI-driven support, we introduced biweekly reflection sessions to encourage metacognitive thinking and self-assessment. We scheduled 30-minute reflection slots every two weeks and provided students with prompts to guide their reflections, such as:

• “What were your key achievements in the past two weeks?”
• “What challenges did you face, and how did you address them?”
• “What are your goals for the next two weeks?”

Students wrote 100-word reflections and pasted them into Smart Agent for analysis. The chatbot provided personalised suggestions for improvement. Based on the chatbot’s feedback, students set goals for the coming two weeks and tracked their progress using the visual progress tracker.

Tips for meaningful reflection

Carefully analyse student reflections to identify common themes, questions or areas of difficulty and update the AI’s dataset using relevant keywords and phrases. For example, if many students mention they struggle with paraphrasing, you could add or enhance content related to paraphrasing techniques in the AI’s knowledge base. This way, when students use terms such as “paraphrase”, “rephrase” or “reword” in their interactions with the chatbot, it can provide specific, tailored responses to scaffold their learning.

By continuously updating the AI system with relevant keywords from student reflections, we ensure that the Smart Agent remains responsive to students’ evolving needs and challenges.

By integrating an AI-powered chatbot into our Moodle-based course, we successfully enhanced students’ self-directed learning behaviours. The combination of short-term achievable goals, personalised AI support and biweekly reflection sessions provided students with the tools and motivation they needed to succeed. This approach enhances engagement, improves learning outcomes and equips students with valuable skills for lifelong learning.

Ying Shao and Xue Yao are associate language lecturers in Xi’an Jiaotong-Liverpool University’s School of Languages.

If you would like advice and insight from academics and university staff delivered direct to your inbox each week, sign up for the Campus newsletter.