How YouTube and games move emotion A K-12 affective media engineering curriculum

Students read how YouTube, games, and films are engineered to move emotion, then design their own work. The same three subjects run from grade 3 through high school, online or in person in Apgujeong, in 1:1 and small-group (1:n) formats.

We never measure a student's emotion by machine. No stage of the program uses AI that records or infers a face, a voice, a gesture, or a physiological signal. Emotion response data comes only from self-report the student marks personally.
KakaoTalk consultation Call 02-540-2922

Written by the CIT teaching team | Reviewed by the CIT curriculum team | Last verified: August 8, 2026 | Next review: at the next curriculum revision

We teach intent and measurement, not editing tricks

The power to make media that moves people comes from stating an intent precisely and measuring the result honestly, not from flashy editing. So every course opens on the same question. Who was this scene built to make feel what?

A student who has answered that question does not only consume media. They can read why a particular color and a particular pace were chosen. In their own work, they set those values themselves.

1. Intent

Decide who should feel what.

2. Spec

Write the emotional goal as design values another person could execute.

3. Direction

Have an AI agent build it, then check the result against the spec.

4. Verification

Confirm with self-report from real people and revise on the data.

The same three subjects, in tools matched to the grade

Reading Emotion, Measuring Emotion, Designing Emotion Systems. The three subjects run under the same names from grade 3 through high school. The tools are Scratch for elementary, App Inventor and Scratch for middle school, and Python for high school.

Reading Emotion

Students watch how picture and sound move a mood. They see and hear for themselves why fast cuts differ from slow ones and bright colors from dark ones. In high school they read the models of emotion science, and the debates still open in it, from primary sources and settle on their own position.

Measuring Emotion

Students mark a mood as numbers on an emotion coordinate map. Elementary students plot the emotional line of a story. Middle school students compute means and distances through A/B feel tests that compare two versions. High school students design and run perception experiments such as contrast calculations and cut-pace curves.

Designing Emotion Systems

From an intended emotion, students set design values in color, sound, and timing and have an AI agent build them. Elementary students make a video letter carrying a feeling, middle school students an app directed through a mapping table, and high school students a capstone directed with Python and a repository.

Close-up still of the silent film actor Ivan Mosjoukine
The same expression reads differently depending on the shots placed before and after it. Actor Ivan Mosjoukine, the material behind the Kuleshov effect covered in Reading Emotion. Photo: Le Brasier ardent (1923), public domain.

Three bands, nine courses, 104 sessions

Session counts are the standard run and are adjusted to the student's pace. Which subject a student starts with is decided in the consultation.

Grades 3 to 5 · Scratch

From the emotion map to a first work

It starts with the habit of marking a mood on an emotion map. With Scratch blocks and a teacher-mediated agent, students carry it through to a first finished work.

  • Reading Emotion, 8 sessions: seeing how fast and slow, bright and dark change a mood.
  • Measuring Emotion, 8 sessions: marking a mood as numbers and drawing the emotional line of a story.
  • Designing Emotion Systems, 16 sessions: writing a wish list and making and delivering a video letter.

All AI calls are mediated by the teacher.

Grades 6 to 8 · App Inventor and Scratch

Meeting it again through experiment and number

Students meet the same three subjects again through experiment and number. Builds are run by an agent working from a mapping table, and the teacher confirms every run.

  • Reading Emotion, 16 sessions: reading thumbnail contrast and hooks by experiment, and checking emotion recognition AI against public images.
  • Measuring Emotion, 8 sessions: using emotion coordinates precisely and computing means and distances through A/B feel tests.
  • Designing Emotion Systems, 12 sessions: directing an agent build with a mapping table that sets design values from a target coordinate.

Agent build runs are confirmed by the teacher.

Grades 9 to 12 · Python

Primary sources, math labs, and a capstone

Reading and measuring are grounded in primary sources and math labs, then the capstone is built through Python and repository-based agent direction.

  • Reading Emotion, 6 sessions: reading the models and the open debates of emotion science from primary sources and settling on a position.
  • Measuring Emotion, 6 sessions: designing and running perception experiments such as contrast calculations and cut-pace curves.
  • Designing Emotion Systems, 24 sessions: planning, building, and verifying a capstone against a spec and pass criteria, then presenting it as a demo.

Publishing a work outside the class goes through approval.

The rules that protect students came first

Because these classes work with emotion, we settled what we will not do before anything else. The rules below apply to every band and every session.

Every session runs the same loop

  1. 1. Mark today's moodClass opens with students marking their own mood on the coordinates. The measuring instrument becomes familiar first.
  2. 2. Read the mediaThe class watches one scene together and reads which emotion arises and why.
  3. 3. Intent and specStudents set the target emotion for their own work and write it out as design values.
  4. 4. Agent build and adjustmentAn AI agent builds from the spec, and the student checks the result and adjusts.
  5. 5. Feel testReal people watch it and leave their response as self-report.
  6. 6. RecordStudents record the distance between goal and result, and the one thing to change next.

The question a feel test settles is never "is this good?" but "did the emotion I specified arise in that person?" What is graded is the spec the student wrote, not the sentence they typed at the AI. Diagnosing a failure is the learning.

Frequently asked questions

What is affective media engineering?

It is a curriculum about how YouTube, games, and video move emotion. Students read how a scene is built and measure the response in numbers. The emotion they intend is written into a spec and built into a work together with an AI agent. Tools are matched to the grade: Scratch, App Inventor, and Python.

What grade can a student start in?

Grade 3 and up. The program is divided into three bands: grades 3 to 5, 6 to 8, and 9 to 12. Which subject a student starts with inside a band is decided in a consultation that reviews the student's experience.

Do you analyze emotion from a camera image of the student's face?

No. No stage of the program uses AI that records or infers a student's face, voice, gesture, or physiological signals. Emotion responses come only from self-report the student marks personally. Even in the sessions that cover emotion recognition AI, the practice work uses public images only.

Can a student join with no coding experience?

Yes. Elementary students start with Scratch blocks, and middle school students use App Inventor and Scratch. Students practice writing exactly what they want built before they practice code syntax. For the high school courses, where students run Python themselves, readiness is confirmed in the consultation.

If the AI builds it, what does the student learn?

What is graded is the spec the student wrote, not the sentence they typed at the AI. Writing an emotional goal as values another person could execute, measuring how far the result landed from that goal, and changing one thing on the evidence of data are work the AI cannot do for them.

Do you cover emotions like fear and sadness?

Only inside a consented story or game frame. For elementary and middle school the ceiling is spooky. Designs that hold attention against the viewer's interest, known as dark patterns, are objects of analysis and are never assigned as builds.

What does a student have at the end?

Elementary students finish with a video letter carrying a feeling, middle school students with an app directed through a mapping table plus the record of their A/B feel tests, and high school students with a capstone built to a spec and pass criteria plus a demo presentation. Awards and admissions outcomes are not promised.

How many sessions is the program?

Across the three subjects the standard is 32 sessions for elementary, 36 for middle school, and 36 for high school. Session counts are the standard run and are adjusted to the student's pace and to the consultation.

We start from what the student already loves

In the consultation we go through the YouTube channels and games the student watches and their coding experience, and settle the band and the starting subject together. Free consultation and placement test available.

KakaoTalk consultation Call 02-540-2922

Related programs

RoboMaster Robot AI

For students who would rather point the same build-and-verify loop at a physical robot and computer vision.

AI education program

Start from a subject interest and design a student-owned AI+X project.

All programs

See every CIT course, from first steps to competition and exam preparation.
Consult (02) 540-2922