ChatGPT Edu HQ POC Sharing Session Monday, 31 August | Meeting | Microsoft Teams

 

• ETD Interactive Learning Production: Lawrence demonstrated how ChatGPT EDU converts existing curriculum documents into interactive mathematics lessons, improves legacy resources, automates tagging, and generates scripts, while stressing that human instructional design and technical troubleshooting remain necessary.
  • Document To Lesson: Pre-U Math - Complex Numbers — Algebra, Geometry and Transformations Interactive with Data

• Lawrence described dropping a Word document into an authorised folder and asking ChatGPT EDU to read it and create an interactive lesson. A worksheet on complex numbers and its learning objectives were converted into practice questions and an interactive activity, then returned for subject-expert review before publication to the MOE Library.
•  Primary 6 Mathematics Revision Quiz Full Coverage
• Primary Mathematics:

• For primary mathematics, Lawrence used a public MOE PDF and learning objectives to generate an interactive lesson. They iterated on the output, including requesting a 3D representation instead of a basic 2D diagram and specifying colour and visual choices that the AI would not infer independently.
• Pedagogical Scaffolding: P1 Word Problems (With Model Drawing) (FA-Math) Lawrence explained that AI can code the interaction but does not independently determine the best teaching sequence. They supplied prompts describing how students should receive hints, observe representations, work through meaningful steps, and connect concrete, pictorial, and abstract representations. The resulting lesson was described as usable but requiring further iteration from student feedback.
• Legacy Resource Modernisation: Lawrence used scripts generated by AI to modernise Community Gallery materials and replace lengthy scrolling with page breaks when questions are detected. They also created an interactive alternative to a symbolic Feedback Math Assistant activity so students could manipulate representations and receive more contextual support.
• Automated Tagging: codexSkill/sls-finalize-community-module/playWright at main · lookang/codexSkill · GitHub For the effort to move Feedback Assistant Mathematics resources into the MOE Library, Lawrence used AI-generated scripts to inspect lessons, detect questions, and assign learning-objective or module-level tags. The process is not fully accurate but reduces the hours required for manual tagging.
• Skill Distribution: GitHub - lookang/codexSkill: Shared Codex skills for SLS and other workflows · GitHub Lawrence stores related instructions and scripts in GitHub, including a skill identified as SLS Finalized Community Module and Playwright automation files. Users can clone the repository and run the provided command file to install dependencies and execute automation, although Rahmuna noted that ChatGPT EDU skills still need to be explicitly shared for workspace access.
• Adoption Plan: Community Gallery  The immediate priority is to prepare and publish existing Feedback Assistant Mathematics resources with CBDD before expanding the broader Community Gallery automation. Lawrence said the work requires local troubleshooting because installations and computer behaviour can vary, and Rahmuna proposed documenting common publishing errors and supporting a future workshop.
• Usage Evidence: The primary mathematics lesson had approximately 1,000 views and six likes in a short period. Lawrence said student feedback was limited, while Rahmuna identified the usage level as an initial indicator worth monitoring and proposed continuing the discussion about scaling and gathering end-user responses.