AI Seoul Showcase
Course specification
AI Literacy and Applied Ethics
Learners establish what a generative model does and does not do, then direct one to produce a researched article and a reasoned ethical position. The unit is assessed by examination.
Main project Examination
Hello World
1.1How a generative model produces language by predicting the next word — and where that reasoning breaks down. Students frame an evidence-based question and write a structured prompt, separating what AI can reliably automate from what it cannot genuinely think through.
Deliverable
A structured, presented deck
Produce a slide deck that follows a fixed brief (a set of instructions) and keeps one look across every slide. A structured slide deck, made from one full prompt and presented aloud.
AI and Work
1.2Directing AI to augment human judgement rather than replace it. Students verify every claim against a named source and brief a model to draft from evidence they supply, examining AI’s real effect on jobs and the workplace.
Deliverable
AI and Work
Look into a claim about AI and jobs, check it against evidence, and publish an article that carries ten marks of humanity. A webpage article about AI and jobs. It carries ten identifiable elements the AI model could not have produced by itself, and any numbers on a chart match the source named for them.
AI and Ethics
1.3Contested applications of AI — algorithmic hiring, autonomous weapons, medical liability — examined through six ethical lenses. Students construct and defend a reasoned argument from an assigned position while naming the cost of the verdict.
Deliverable
AI and Ethics, the Tribunal
Argue a real AI case from the point of view you are given, then explain the opposing side fairly. A written legal document that argues the point of view you were given, plus a spoken argument in front of the judges, who then give their decision.
The Test
1.4Consolidating and assessing the module’s knowledge under examination conditions. Students apply spaced revision, build a study tool that forces active recall, and sit a timed, closed-book test on the module’s content.
🏆 Capstone deliverable
The Test
Test how well you can recall facts, by building your own study game, then sit the exam. A study game that runs in a browser and makes the player recall the content from week one. Learners also sit a timed test that counts towards their grade, on the course platform.
CAPSTONE
The Test
AI Media Production
Teams produce a film trailer built on a collision between the culture they arrive with and the culture they are working in. Structure is fixed in writing before any asset is generated.
Main project Film — Six crew roles produce one trailer for a film that does not exist, built on a collision between the culture the team arrives with and the culture they're working in. The plan, a logline and a six-beat spine, is locked in writing before a single image is generated. From there the team locks a style bible, generates and curates keyframes and motion, cuts to a scored track, and finishes the film to a delivery spec. It premieres at a live screening: four to five minutes, the whole trailer, a pitch, and a fixed critique from the room.
The Plan
2.1Planning a film trailer in writing before any image is generated. Students write a logline and a six-beat structure, then use a model to expand and refine story ideas by hand into a text-only plan.
Deliverable
The Plan
Fix the trailer's logline and six-beat structure in writing before any pictures are created. A text-only plan: one logline sentence, plus a sheet listing the six beats.
Image Generation
2.2Specifying a film’s visual style in words and generating a consistent image sequence from it. Students define aesthetic, lighting, mood and palette, train a reference for character consistency, and control drift and cost across a twelve-image spine.
Deliverable
Image Generation
Produce twelve images that look like they belong to the same film, forming the visual spine of the trailer. A set of twelve final-quality images, all in one fixed aspect ratio, covering the world of the film and the main character.
Editing to Sound
2.3Editing picture to sound and acquiring the vocabulary of camera direction. Students generate a music track to a chosen mood, cut the picture to its beat, and chain a text model into an audio model to produce narration and score.
Deliverable
Editing to Sound
Make a section of the trailer with music, where the picture is cut to match an AI-generated music track. A piece of music made to match a mood and cut together with the picture, plus a task list where every job has one named person responsible.
Animating Stills
2.4Turning still images into controlled motion, and diagnosing failed generations. Students read camera movement from real trailers, write four-part motion prompts, and cull a large set of clips down to the few that hold together.
Deliverable
Animating Stills
Turn a still image into a short moving clip with controlled movement. Name any problem before you change the prompt to fix its cause. A set of short moving clips. Each one is saved with its four-part motion prompt and a written review.
The Oscairs
2.5Assembling, presenting and evaluating a finished trailer. Students cut to the soundtrack, build a one-page promotional site, and defend their creative choices at a live premiere judged against a shared evaluation spine.
🏆 Capstone deliverable
The OscAIrs
Show a finished trailer and its promotional website at a premiere, and explain every choice made using an AI tool. The six items brought to the stage, including the finished trailer (edited to fit the soundtrack from topic 2.3) and a one-page promotional website, plus one shot posted as before and after with its instructions and an evaluation.
CAPSTONE
The OscAIrs
Six crew roles produce one trailer for a film that does not exist, built on a collision between the culture the team arrives with and the culture they're working in. The plan, a logline and a six-beat spine, is locked in writing before a single image is generated. From there the team locks a style bible, generates and curates keyframes and motion, cuts to a scored track, and finishes the film to a delivery spec. It premieres at a live screening: four to five minutes, the whole trailer, a pitch, and a fixed critique from the room.
AI Product Development
Teams define a product, deploy it, ground an agent in a database they build, chain live APIs into a working service, cost it, break it deliberately, and pitch it. Learners hold named company roles throughout.
Main project App
Business Model Canvas
3.1Defining a product with the Lean Canvas and taking it online. Students map an idea across the nine boxes from a real customer problem, write a value proposition, and use an AI coding tool to build and publish a landing page.
Deliverable
Business Model Canvas
Turn a product idea into a completed canvas and a live landing page, in one evening. A completed nine-box business model canvas, plus a landing page published to a public web address.
Page Audit / API
3.2Auditing a live page for unsupported claims and grounding an agent in your own content. Students score a page against five measures, remove unsupported “slop”, and build a password-protected agent that answers only from supplied material.
Deliverable
The API
Check a live website against five measures. Then build a working agent (chatbot) page that only talks about your own content, fixed using real error messages, and hand it in by 10pm. A live agent (chatbot) page: a working web address, protected by a password, that can hear spoken questions and reply out loud, and only answers about your own content. The secret key is kept in the host's environment variables (its settings area). Plus one written website check.
App Development
3.3Designing an agent persona grounded in a structured database. Students model tables and scoped access keys, write a persona as a job description, and build an AI agent that answers only from a supplied catalogue.
Deliverable
App Development
Build an AI agent that uses a structured database, so it only answers using facts from a supplied catalogue. A designed database (four tables with example rows of data and a stated type for each column), an access key that only allows certain actions, rows of data written by Codex and checked one by one, and an AI agent that pulls in this data. A completed persona plan on paper.
Hackathon
3.4Shipping a working, multi-page product against a deadline in role-based teams. Students research a comparable company, diagram how code, database, AI and APIs connect, and deliver a landing page, a live-data gallery and a grounded agent.
Deliverable
The Hackathon
Turn an idea into a working website with a database and an AI agent, and get it online before the deadline. A three-page website (a landing page, a gallery backed by a database, and a password-protected AI agent), plus a portfolio of supporting work: the Lean Canvas (a one-page business plan), the published site, the AI agent, the database, a diagram of how it is all built, the cost workings, and the list of problems found and fixed (the defect log).
The Final Sprint
3.5Testing a product adversarially, remediating and pitching it. Students hunt defects, red-team their own agent, keep a public defect log, and deliver a costed live pitch to a room that did not build the product.
🏆 Capstone deliverable
The Final Sprint
Find defects in your own product and agent before users do, then pitch the product and show it working live. A public defect log, plus a five-minute pitch on four pillars, with the product shown working live and a Codex-generated deck, infographic and promo pack, all cut to the four pillars.
CAPSTONE