Tools change every few months; what we ask them to do does not. This session maps AI tools for research by use (compressing, extracting, expanding, formalising) and by the kind of data they may legitimately receive under UZH rules, including where the university’s own tools are and how to access them. Participants triage six real research tasks, then watch tools at work in a live carousel, which includes a short introduction to the Model Context Protocol, the standard that lets AI tools talk to each other. The core exercise is critical evaluation done as a procedure, not an attitude: groups verify, claim by claim, the answers three different AI tools gave to the same literature question, and measure where each one fails. The morning ends with each group stress-testing a research workflow for its own question, the basis of the final presentation.
