AI for Article Writing: a Soft Toolbox

April 1-2, 2026

Debora Zoia, CUSO MPS

Dr. Romain-Daniel Gosselin, DataBio

This immersive two-day course aims at equipping participants with knowledge and practical skills in AI-assisted scientific writing. The program covers AI fundamentals, ethical considerations, and hands-on applications. Participants will learn good practices as well as major caveats in prompt engineering, scientific writing, and literature retrieval using usual web-based AI tools. Through interactive workshops and problem-solving sessions, students will develop critical skills, all without requiring technical expertise in data science.The course particularly emphasizes the responsible use of AI in research.

Day 1: Foundations and practical AI applications in writing • Introduction to AI in life sciences: Brief history of AI/ML/NLP/LLM; introduction of algorithm functioning; current landscape of available tools for life scientists. • Ethical considerations in AI: Research integrity; potential biases; considerations about authorship and citations; data privacy and confidentiality; institutional guidelines. • Prompt engineering fundamentals: CraNing effective prompts; zero-shot vs few-shot learning; techniques for scientific writing; avoiding and spoPng allucinations; practical exercises in prompt design. • AI for scientific writing: Literature retrieval and summarization; draNing manuscript sections; reference management; edition and refinement techniques; hands-on writing workshop. • Practical session: Writing: Use of generic (ex: ChatGPT, Claude) and specific (ex: askyourpdf, consensus) tools; comparative analysis of outputs and capabilities; Hands-on exploration.

Day 2: Advanced AI functionalities of AI in the management of scientific contents • AI in literature retrieval: Use of generic (ex: ChatGPT, Claude) and specific (ex: askyourpdf, consensus) tools. • PiXalls of AI use in article writing: Be wary of 'hallucinations'; issues with content standardisation and tortured phases. • Practical workshop: writing: Group challenges: writing a short (2 pages) referenced article on a given topic (free choice of tools); presentations of group outputs; critical evaluation. • Can I peer review helped by AI?: Extracting information from a manuscript; detecting the use of AI; structuring a review report; journal regulations. • Practical workshop: peer reviewing: Group challenges: peer reviewing a given (short) text; focus on the detection of AI contents, detection of fabricated images, and the writing of the review report. • Caveat with AI detectors: Presentation of detectors; AI detectors lack reliability; A variety of outputs. • Final conclusions: Course wrap-up, perspectives, recommendations for continued learning. 

UNIL

Registration on the CUSO MPS webpage about 2 months before the activity starts.

3