Title: Magic Mirror – Generative Image AI, Body Images, and Power
Date: February 11, 2026, 12:00–03:30 PM
Type: Workshop/Seminar, Presence
Location: Foyer des Augusteums
Event series: “Action Day” Women in Science 2026 – Alliances for Change
Project: “Who designs for whom in AI?” What AI has in common with power, feminism and a (distorted) mirror

Generative image AI often appears to be an objective tool – but behind the images lie social norms, prejudices and reproduced power structures. With the interactive AI project “Magic Mirror”, we want to make this dynamic tangible and open to discussion: participants recognise in their owen ‘mirror image’ how AI makes decisions, what attributions arise and how fairness is negotiated. 

We invite you to join us in taking a critical look at the topic of ‘Generative Image AI, Body Images and Power’ and to contribute your personal experiences. An interactive mirror installation, the ‘Magic Mirror’ makes AI pipeline steps visible, algorithmic decisions comprehensible and problematic outputs discussable. Through practical exercises on prompt sharpening and targeted perspective shifting, you will learn how attributions arise, how to critically address them, and how to strengthen AI literacy in a targeted manner. Feminist perspectives and moderation ensure that all experiences are reflected upon respectfully and safely. The aim is to promote and discuss responsibility, judgement, and the ability to act when dealing with generative AI.

This Hands-On Workshop is part of the action day „Women in Science 2026 – Alliances for Change”, the International Day of Women and Girls in Science at Leipzig University. Further information is available here. The workshop is organized in cooperation between Vanessa Kuhfs und the team of CitizenScience.AI at ScaDS.AI Dresden/Leipzig.

Speakers

Vanessa Kuhfs
Design Research & FeministAI
Universität Leipzig, ScaDS.AI Dresden/Leipzig
Kunsthochschule Burg Giebichenstein Halle
vanessa.kuhfs(at)uni-leipzig.de

Johannes Häfner
Data Scientist, R&D Engineer
Universität Leipzig, ScaDS.AI Dresden/Leipzig
johannes.haefner(at)uni-leipzig.de