제주특별자치도 공공정책연수원
Innovation leader program — three-stage AI course
About the program
The innovation leader program is not a one-off lecture. The same cohort works through three stages over several months. Stage one builds the basic fitness for working with AI, stage two produces a real deliverable in the form of a provincial promotional video, and stage three analyzes the data behind policy decisions. Stage one covered prompt engineering, image AI, video storytelling, and AI ethics over two days, setting up the tools and judgment the later stages rely on. Stage two spent four days finishing one promotional video from start to end: writing the storyboard, turning still images into video, generating and placing background music, then cutting and subtitling. Each participant had dedicated working time, and the instructor and assistant instructor resolved blockers at the desk as they came up. Stage three works with the large datasets that policy decisions rest on. Material that includes open-ended survey responses is organized with basic Excel formulas, then carried through to statistical analysis and dashboard building with AI. The goal is that an innovation leader who completes all three stages can produce promotional material and data reports in their own department.
What we ran
A program in which the same cohort moves through three stages over several months: AI fundamentals, promotional video production, and data analysis
How the course was designed
Each stage leaves one deliverable behind: usage standards in stage one, a finished promotional video in stage two, a data dashboard in stage three
Tools learned in one stage are used directly in the next, so the learning carries forward
Practice material drawn from real Jeju administration, such as provincial promotion and policy data
Main components
[Stage 1] AI fundamentals
Prompt engineering, image AI, video storytelling, and AI ethics over two days
[Stage 2] Producing a provincial promotional video
Storyboard → still images to video → background music generation and placement → cutting and subtitles, through to one finished video
[Stage 3] Data analysis and visualization
Organizing large policy datasets, including open-ended responses, with basic Excel formulas, then statistical analysis and dashboards with AI
How it ran
Multi-session classroom training by stage: two days for stage one, four days each for stages two and three
Dedicated working time in every session, with the instructor and assistant instructor supporting at the desk
Video editing software, headsets, and the rest of the practice environment prepared in advance
Cautions covered
Copyright and likeness rights in generated images and video, and the permitted use of provincial material
Rules on never entering personal data or non-public administrative records
Gallery
