A hands-on workshop for students and researchers of any field and any level who want to use AI in their research in a safe, ethical and effective way: to work faster and to improve their ideas, without handing over the thinking that makes the work their own. 

In three and a half hours you will take a research idea of your own, shrink it to a size you can build, build it with an AI assistant, fix what goes wrong, run it, and analyse your data with an analysis you planned in advance, ending with a report of your results (in R, another tool of your choice, or directly in your browser). Experiments, questionnaires, simulations and existing public datasets are all supported, and two fully worked examples show every step. 

You will learn how to: 

• Turn your own research idea into a study you can build and run in an afternoon; 

• Recognise the stages of a research project where AI genuinely helps, and those where it quietly replaces your own judgement; 

• Choose the right tool and plan for the task (Claude and Claude Code, ChatGPT and Codex, Gemini and Antigravity, and Microsoft Copilot Chat, which Unifr provides), including what the free versions can and cannot do;

• Write prompts that work as specifications rather than wishes; 

• Use AI to build, debug and test an experiment or questionnaire, and to write an analysis before collecting any data, tested on made-up data; 

• Check AI output, including code that runs without any error and is still wrong; 

• Use AI with academic integrity: what it should not write for you, and how to disclose its use. 

About 60% of the time is hands-on, with dedicated time for debugging and for your questions. Participants who have never written a line of code are explicitly welcome: the session is built so that they succeed. 

Objectifs

The workshop addresses all six levels, with most of the time spent on Apply, Analyse, Evaluate and Create. In the terms of the infographic below, AI is used for what it does well (suggesting alternatives, drafting code), while participants practise the distinctive human competences: critical judgement, verification of results, ethical reflection and original design. 

• Remember: Identify the three kinds of AI tools used in research (chat assistants, literature tools, coding agents) and the stages of a research project;
• Understand: Explain in plain words how a large language model produces text, and why it can state false information with full confidence;
• Apply: Write a prompt that works as a specification and use it to build and run a working experiment or questionnaire (PsychoPy, Python, MATLAB or plain HTML);
• Analyse: Find errors in AI-generated code that produce no error message, and analyse their own data with a pre-planned analysis, then read the resulting report critically;
• Evaluate: Judge an AI-suggested design for novelty, confounds and fit; test a peer's study by trying to break it; decide which parts of a report AI may help with, and how to disclose its use;
• Create: Design an original small study, produce real data and a results report from it, and leave with a personal protocol for using AI in their own thesis. 

 

Public-cible

Etudiant·e·s

• Bachelor students starting to plan a thesis, who want good habits from the start;
• Master students in the middle of a thesis, already using AI without a clear sense of where it belongs;
• Doctoral and postdoctoral researchers who want to accelerate experiment development, coding and analysis without compromising rigour
• Students with no programming background who have been told that coding their own study is out of reach. A central aim of the session is to show that this is no longer true;
• Anyone uneasy about the ethics of AI in academic work, who would rather resolve the question with clear principles than avoid the tools. 

Prérequis

• Basic computer literacy: comfortable using a laptop, a web browser, and managing files;
• Access to a general-purpose AI assistant (see Material Needs for details on how this is handled);
• No programming experience is required. No prior experience with AI is required. 

Responsables et intervenants

Intervenant(s)

Gabriel Leipner, final year of Master’s in Clinical Neuroscience

Dates et lieux

Période	Lieu
16.11.2026 de 13:15 à 17:00	B205