DHUM 25A43 - Investigating with AI

How can we leverage cutting-edge AI tools to produce digital investigations for social sciences and political analysis?

This course teaches students to use AI — Large Language Models and agents — for data collection, analysis, and visualization, while critically assessing its strengths and limitations. It covers the foundations needed to work confidently with these tools: NLP, language models, APIs and data formats, data visualization, and the sovereignty and environmental questions they raise.

The heart of the course is a hands-on group investigation. Students choose a topic, design a research protocol, and analyze real-world data — parliamentary speeches, climate reports, social media discourse — then present their findings as a website or live presentation. Each student also writes a reflective essay on their experience working with AI.

No prior technical knowledge or coding experience

The course runs over twelve sessions that alternate short lectures and live demonstrations, leaving ample room for hands-on practice. Each session introduces a concept — an AI tool, a data source, a method of analysis — and shows it in action on a real example. Students work in groups on their investigation project throughout the course, and receive close support from the instructor to design their approach, navigate the tools, and refine their analysis. The final sessions are dedicated to project work and presentations. No prior technical background is required — curiosity and a willingness to experiment with AI tools are what matters

Students are evaluated through two complementary assignments that together meet the requirement of varied assessment methods, combining group work with individual reflection, and in-class participation with out-of-class production.

• Group investigation project (collective). Working in groups of two to three, students carry out a data journalism–style investigation on a topic of their choice. They identify a dataset, define a research question, conduct the analysis using the AI tools and methods seen in class, and present their findings at the final session. Group work is developed progressively across sessions, with regular checkpoints and feedback during class. Evaluation criteria include the rigor of the investigation, the quality of the analysis, the clarity of the storytelling, and the soundness of the methodological choices.
• Individual reflective essay. Each student writes a short individual essay reflecting on their experience conducting research with AI throughout the project — what worked, what didn't, where AI helped and where it got in the way, and what they learned about doing investigative work alongside these tools. This essay is submitted at the end of the course and assessed individually.