AI Summercamp 2025

Techniques and Ethics of Generative AI

Author

Dr. Brian Ballsun-Stanton (Macquarie University)

Published

September 8, 2025

Welcome

Welcome to the AI Summercamp 2025 at Justus-Liebig-Universität Gießen!

This three-week intensive course explores the epistemological boundaries between what students believe LLMs are versus what they actually are. Human judgment remains central throughout.

Basic Course (8-12 September)

This intensive week introduces AI literacy through hands-on exploration, systematic annotation, and progressive skill building. Each day addresses a fundamental question about working with Large Language Models.

Each day consists of:

Morning session (9:00-10:30): Skills development
Break (10:30-11:00): Grimoire work and annotation
Midday session (11:00-12:30): Critical discussions and judgment development

Day 1 (Monday, Sept 8): Can we control AI output?

Session 1: Welcome and Getting Started - Setup, introductions, goal setting
Session 2: The ‘Confident’ Assistant - System prompts, drunk tutor problem

Day 2 (Tuesday, Sept 9): How do we populate the context window?

Session 1: Basic Prompting Foundations - Rules of thumb, annotation practice
Session 2: Metaprompting - Blank page problem, epistemic humility

Day 3 (Wednesday, Sept 10): What can we verify?

Session 1: File Handling and Source Work - Extraction vs interpretation
Session 2: Model Differences - Service grain, failure patterns

Day 4 (Thursday, Sept 11): How should we work?

Session 1: Research Infrastructure - State management, conversation strategy
Session 2: The Boring Important Stuff - Legal, ethical, privacy discussions

Day 5 (Friday, Sept 12): When do things break?

Session 1: Breaking Things Systematically - Failure modes, evaluation criteria
Session 2: Synthesis and Next Steps - Reflection, community building

Core Learning Objectives

By week’s end, participants will:

Understand fundamental concepts - Context windows, tokens, confidence paradox
Develop practical skills - Prompting, verification, state management
Recognize limitations - Confabulation, model differences, failure patterns
Build judgment - When to use, when to avoid, how to verify

The Grimoire

Throughout the week, we build a collective “spellbook” of prompts through:

Daily “Show me your prompts” sessions
Systematic annotation practice
Shared documentation on Conceptboard
Peer learning from successes and failures

Advanced Course (15-19 September)

Time: Monday-Friday, 09:00-12:30
Location: Hauptgebäude, Seminarraum 315
Language: English
Instructor: Dr. Brian Ballsun-Stanton

Week Overview:

Day 1: Anthropic Documentation & Coding Assistance
Day 2: Test-Driven Development & Transcription
Day 3: Embeddings & Systematic Literature Review
Day 4: SillyTavern & Corpus Linguistics
Day 5: Project Work & Crumple Zone

Research Week (22-26 September)

Individual consultation on AI in research with the summer camp team.

Contact

Dr. Brian Ballsun-Stanton
Macquarie University
📧 brian.ballsun-stanton@mq.edu.au

--- title: "AI Summercamp 2025" subtitle: "Techniques and Ethics of Generative AI" author: "Dr. Brian Ballsun-Stanton (Macquarie University)" date: "2025-09-08" toc: true --- ## Welcome Welcome to the **AI Summercamp 2025** at Justus-Liebig-Universität Gießen! This three-week intensive course explores the epistemological boundaries between what students believe LLMs are versus what they actually are. Human judgment remains central throughout. ## Basic Course (8-12 September) This intensive week introduces AI literacy through hands-on exploration, systematic annotation, and progressive skill building. Each day addresses a fundamental question about working with Large Language Models. Each day consists of: - **Morning session** (9:00-10:30): Skills development - **Break** (10:30-11:00): Grimoire work and annotation - **Midday session** (11:00-12:30): Critical discussions and judgment development ### Day 1 (Monday, Sept 8): Can we control AI output? - [Session 1: Welcome and Getting Started](basic/day1-1.qmd) - Setup, introductions, goal setting - [Session 2: The 'Confident' Assistant](basic/day1-2.qmd) - System prompts, drunk tutor problem ### Day 2 (Tuesday, Sept 9): How do we populate the context window? - [Session 1: Basic Prompting Foundations](basic/day2-1.qmd) - Rules of thumb, annotation practice - [Session 2: Metaprompting](basic/day2-2.qmd) - Blank page problem, epistemic humility ### Day 3 (Wednesday, Sept 10): What can we verify? - [Session 1: File Handling and Source Work](basic/day3-1.qmd) - Extraction vs interpretation - [Session 2: Model Differences](basic/day3-2.qmd) - Service grain, failure patterns ### Day 4 (Thursday, Sept 11): How should we work? - [Session 1: Research Infrastructure](basic/day4-1.qmd) - State management, conversation strategy - [Session 2: The Boring Important Stuff](basic/day4-2.qmd) - Legal, ethical, privacy discussions ### Day 5 (Friday, Sept 12): When do things break? - [Session 1: Breaking Things Systematically](basic/day5-1.qmd) - Failure modes, evaluation criteria - [Session 2: Synthesis and Next Steps](basic/day5-2.qmd) - Reflection, community building ### Core Learning Objectives By week's end, participants will: 1. **Understand fundamental concepts** - Context windows, tokens, confidence paradox 2. **Develop practical skills** - Prompting, verification, state management 3. **Recognize limitations** - Confabulation, model differences, failure patterns 4. **Build judgment** - When to use, when to avoid, how to verify ### The Grimoire Throughout the week, we build a collective "spellbook" of prompts through: - Daily "Show me your prompts" sessions - Systematic annotation practice - Shared documentation on Conceptboard - Peer learning from successes and failures ### Advanced Course (15-19 September) **Time:** Monday-Friday, 09:00-12:30 **Location:** Hauptgebäude, Seminarraum 315 **Language:** English **Instructor:** Dr. Brian Ballsun-Stanton **Week Overview:** - **Day 1:** Anthropic Documentation & Coding Assistance - **Day 2:** Test-Driven Development & Transcription - **Day 3:** Embeddings & Systematic Literature Review - **Day 4:** SillyTavern & Corpus Linguistics - **Day 5:** Project Work & Crumple Zone ## Research Week (22-26 September) Individual consultation on AI in research with the summer camp team. ## Contact **Dr. Brian Ballsun-Stanton** Macquarie University 📧 [brian.ballsun-stanton@mq.edu.au](mailto:brian.ballsun-stanton@mq.edu.au)