Training models, building agency.

Overview

The AI Ambassadors workshop introduces participants to generative AI not as a finished product to be consumed, but as a technology they can shape and direct. Over three days, participants train their own small-scale language models, design AI agents for professional scenarios they envision for their own futures, and reflect on the relationship between humans and these tools as well as the ethical issues they raise. The workshop is part of the ATOM (Age Against the Machine) project at the University of Oulu, which investigates how hands-on AI literacy activities can support young people's sense of agency in shaping their own working lives in an era of rapidly changing technology.

📂 CategoryInnovation
📌 KeywordsAI literacy, future of work, Large Language Models, Machine Learning, AI agents, human-in-the-loop
🎯 ObjectivesBy the end of the workshop, participants will:
 
1) Understand the basic mechanics of how language models are trained and how their behaviour is shaped by training data
2) Gain hands-on experience designing and prototyping LLM-based agents
3) Develop a more grounded sense of personal agency in relation to AI technologies 
🌐 LanguageEnglish (Japanese support available)
👥 Target ParticipantsYouth (aged ~15-25)
🗓 FormatSingle session（4-hour in-person）
✅ Past & Upcoming Sessions#1; Sunday, October 18, 2026, 13:00-17:00
#2; Sunday, October 25, 2026, 13:00-17:00
📍 VenueS4-202, South Building 4, 2nd Floor, Ookayama Campus
🌟 Recommended for• Students who are curious about how AI systems work
• Students thinking about their future careers in the AI-shaped workplaces and industries
• No technological or programming background required
📊 Points8 points (4 points x 2 sessions) (Points to be awarded to students enrolled in the courses for credits. Please refer to the syllabus for details.)
🎟️ Capacity20 participants

Main Content

(Flexible and subject to change according to practicalities)

Part 1 — Foundations & Model Training: Introduction to the practicalities of research. Brief introduction to language models and how they learn. Hands-on pre-training and fine-tuning of small-scale language models using the Little Language Machine educational tool. Participants observe directly how training data shapes model behaviour.

Part 2 — From Models to Agents: Introduction to LLM-based agents and the concept of human-in-the-loop systems. Participants explore future professional scenarios relevant to their own interests and begin designing and implementing AI agents to operate within them, making decisions about how humans and the agent should share responsibilities.

Part 3 — Implementation, Review and Reflection: Continued implementation and peer review of agent prototypes designed on Day 2. Reflection on the broader implications of AI on the future of work and society.

Facilitator

Pauli A. Klemettilä (クレメッチラ・パウリ)
INTERACT Research Group, University of Oulu

Pauli A. Klemettilä is a PhD candidate at the INTERACT Research Group, University of Oulu, studying AI literacy with young people. His prior work includes research on AI perceptions among Finnish and Japanese children and youth, with publications at CHI 2025 and IDC 2025. Additionally, his master's thesis on critical AI literacy with children won the SIGCHI Finland Master's Thesis Competition in 2025.

Outside of research, he practices kendo at a local club in Oulu.

Organizer

Seiya Moro
2nd year doctoral student

Graduate Major in Engineering Sciences and Design, Department of Transdisciplinary Science and Engineering, School of Environment and Society, Institute of Science Tokyo

・First year master’s course student, Department of Technology and Innovation Management (Dual Degree)
・Research Fellow, Japan Society for the Promotion of Science (JSPS)

Research Topic:
"Developing a self-care independence support system for children with severe intellectual disabilities and their families through co-design."

Contact Faculty

Yuri Matsuzaki
📩 matsuzaki.y.77a9 [at] m.isct.ac.jp