Section 1: AI in the Public Sector: Opportunities and Boundaries
- What generative AI does well, and which tasks should never be left to AI instead of an officer
- Tasks that benefit: drafting letters, summarising meetings, searching regulations and answering citizens
- Limits to know: wrong information (hallucination), bias and out-of-date knowledge
- Accountability: the officer checks and is responsible for every piece of work that goes out
Section 2: Lab: Data Classification and PDPA Before Using AI
- Classifying data: public, internal use, personal and confidential
- PDPA principles officers should know when using AI with citizens' data
- Personal versus agency accounts, chat history settings and whether data is used to train models
- Techniques for masking personally identifiable information before sending text to AI
- Lab: sort sample documents by data class, then mask the personal data
Section 3: Lab: Prompting for Government Work
- Prompt structure: role, context, task, output format and constraints
- Setting the right register for official Thai and for communicating with the public
- Adding examples and asking the AI to ask back when information is missing
- Lab: compare ChatGPT, Gemini and Copilot on the same task
Section 4: Lab: Drafting Official Letters and Memos
- Draft internal memos, external letters and replies to citizens from short briefs
- Use the agency's own templates as examples to get the correct letter structure
- Have AI check language, concision and politeness
- Always double-check names, positions, figures, dates and any references AI may invent
- Lab: draft a letter from a sample brief, then review it with a checklist
Section 5: Lab: Meeting Summaries and Minutes
- Transcribe with tools the agency allows, and what to watch for when recording meetings
- Summarise into minutes: issues, resolutions, owners and deadlines
- Extract follow-up actions and draft a circular reporting the meeting results
- Lab: turn sample meeting notes into minutes using a standard form
Section 6: Lab: Searching and Summarising Regulations with NotebookLM
- Build a notebook from public regulations, operating manuals and circulars
- Ask questions with cited sources and always open the original passage
- Summarise into comparison tables, FAQs and executive briefs
- Lab: build a knowledge base of operating manuals from sample documents