Introduction to AI @SciencesPo Le Havre
This six-hour course offers an accessible yet rigorous introduction to artificial intelligence for students with no technical background. Through live demos, conceptual explanations, and hands-on practice, we move from the foundational mechanics of large language models to the broader questions they raise about work, intelligence, the environment, and geopolitical power. The course is designed to make you a confident, critical, and effective user of tools that are already reshaping the world you will work in.
Session 1 – Aug 31 2026, 2-5 pm
Intro & Demos (15mn)
The AI landscape (2h)
We present a brief history of AI and where we are now focusing on an acceleration that has few historical precedents. We present the landscape of models in all their diversity (size, open source, use cases etc) and show how very specific benchmarks are needed for comparisons. LLMs form the foundation of a rapidly expanding ecosystem from autonomous agents to voice AI and synthetic personas. We introduce a selection of powerful AI tools.
Society (1h30)
We address some of the most pressing questions related to the rise of AI, including its environmental impact and the issue of workforce replacement. We consider the question of the intelligence of AI and the phenomenon of emergence as unpredictable jumps in capability that arise as models scale. And we look into AI sovereignty based on recent events.
Session 2 – Sept 1st 2026, 9-12 AM
How it works, the technical bits (45mn)
Through a conceptual overview, we build an intuition for how LLMs work and address the foundations: embeddings and semantic similarity and the notion of tokens. This leads to models training, architecture and complexity. Students have a high level understanding of how AI works
Work with AI (1h15)
This last section guides the students into efficient and qualitative work with AI. The apparent chatbot simplicity lures us into passivity and total delegation. Genuine work with AI is, by contrast, intense as it relies on a high level of cognitive abilities. Concretely we look into prompting techniques, working methods and the studies and concepts on which they are based. We explore the more advanced features of Claude and ChatGPT.