AIED — AI in Education

Path: AIED/

A scholarly review article (target: 20-40 pages) on AI in education, analyzing how AI redistributes feedback, attention, judgment, and institutional power. Swedish is the canonical source language; an English adaptation is derived from it.

Scope

The article examines AI in education through four analytical lenses:

1. Feedback — how AI-mediated feedback changes the learning loop
2. Attention — how AI shapes what (and who) gets attention in educational settings
3. Judgment — how AI systems exercise judgment that was previously human
4. Institutional power — how AI redistributes authority in educational institutions

Structure

AIED/
├── article/
│   ├── sv/ai-i-utbildning.qmd    — Swedish source manuscript
│   └── en/ai-in-education.qmd    — English adaptation
├── plans/                         — Planning docs
│   ├── agent-coordination-board.md
│   ├── philosophical-mapping.md
│   └── implementation-plan.md
├── _quarto.yml                    — Quarto config (HTML/PDF/GFM)
├── Dockerfile                     — Reproducible build
├── docker-compose.yml
├── Makefile
├── META.md                        — Project state meta-analysis
├── FEEDBACK.md                    — External review of the plan
└── TECH.md                        — Technical build instructions

Technology Stack

Layer	Technology
Publishing	Quarto
PDF Engine	XeLaTeX
Diagrams	Mermaid
CSS	Tufte-inspired styling
Fonts	Noto family
Build	Docker + Makefile
Languages	Swedish (source), English (derived)

Build

# Build all formats
make

# Build specific format
make html
make pdf
make gfm      # GitHub-flavored markdown

# With Docker (reproducible)
docker compose run --build article make

Coordination

The project uses a multi-agent coordination protocol (detailed in AGENTS.md) for async writing work. An agent coordination board tracks progress across:

• Manuscript writing (SV and EN)
• Evidence localization
• Production build
• Delivery package
• Consistency and proofing passes

Current Status

The backbone is complete. Evidence localization has been done for both Swedish (sections 3-14) and English (sections 1-14). Production build, delivery package, consistency pass, diagram SVG assets, typography pass, and philosophy audit are all complete.