eMPF — Project Roadmap

1Identify a government decision▶

Choose a concrete decision (e.g. eMPF platform design, user onboarding, data portability). Align with SDGs.

AI can help: Search Legco/gov.hk for eMPF, MPFA, pension; list meetings and papers; draft candidate decisions.

Human steers: Decide which decision to investigate; confirm scope and course alignment.

2Review the decision — room for improvement▶

Document how the decision is controversial or disputable (e.g. adoption data, usability). Frame as room for improvement.

AI can help: Summarise Legco/news on eMPF; suggest gaps; draft an issues list.

Human steers: Choose which gaps to focus on; set tone and scope.

3Search and curate public information▶

Collect gov and Legco docs on eMPF. Focus on decision-making and data management. Use government telephone directory.

AI can help: Search/fetch links; extract quotes or tables; organise findings; suggest contacts.

Human steers: Choose authoritative sources; verify facts; decide what to cite.

4Raise questions under the Code on Access to Information▶

Draft clear request to relevant body (e.g. MPFA). Send by Weeks 3–4; expect ~3 weeks (up to 7) for response.

AI can help: Draft request letter; suggest Code citation; check clarity.

Human steers: Finalise and send request; manage follow-up; interpret response.

5Collect community data (in parallel)▶

Design small-scale collection (e.g. survey on eMPF adoption, usability). Use fieldwork allowance (HK$300/student; Week 6 deadline).

AI can help: Draft survey template; suggest tabulation; help build CSV.

Human steers: Design and conduct fieldwork; interpret data in context.

6Analyse data and make recommendations▶

Use math and visualization (stats, comparison, Python/Jupyter). Link every chart to a question. Produce recommendations.

AI can help: Write/run code; generate charts and stats; draft bullet recommendations.

Human steers: Choose analyses; interpret and write narrative; decide what to present to Legco.

7Present to Legco via the Redress System▶

Complete report (~3,000 words), poster, two presentations. Frame as complaint on data governance. Submit via Legco Redress System.

AI can help: Draft report sections; suggest structure; format poster/slides.

Human steers: Set argument and tone; ensure accuracy; present and submit.