Images shape how we understand science even before we might have read a single sentence. They tell us what research looks like, where it happens, and, perhaps most importantly, who does it. On October 6th, 2026, I had the opportunity to explore these questions with participants at the University of Ottawa during a workshop titled Reflecting on Images in Science – Using AI Responsibly and Creatively, hosted by the Centre for International Policy Studies (CIPS).
Why Look at Images at All?
In science communication, images are often treated as decoration: something to make a press release or a website look more attractive. Yet every image is a decision. It shows one thing and leaves out another. It suggests a mood, a hierarchy, a norm. When the same visual patterns repeat over and over, the white lab coat, the bubbling flask, the lone genius at the blackboard, they quietly define what we expect science to be.
The starting point for the workshop was the Bildreflektor (image reflector), a tool for the systematic analysis of visual bias that I developed within the “Siggener Kreis” (Siggen Circle), a German high-level working group on science communication. The idea is simple: instead of judging images quickly and intuitively, we slow down and look at them together, with a set of questions that make our own assumptions visible.
Looking Closely at Images
Working with printed examples spread across the tables, participants sorted images from science communication in two rounds. In the first round, they arranged them by emotionality, from highly emotionally charged images to very factual ones. In the second round, the focus shifted to otherness: from the stereotypical image to the original one. The aim of this first exercise was to sensitize participants for the different messages that images can convey and get a conversation about their functions going.
What makes this exercise valuable is not finding the “right” answer. It is the moment when a group notices that they read the same image very differently, depending on their discipline, background and experience. One insight emerged clearly from the discussion: the same image can be perceived in completely different ways, and there is no single image that works for every purpose. For instance, a picture of a young, female chemist with coloured hair provoked very different responses, from “not scientific at all” to “finally a break of stereotypical misperceptions”.
Missing Pictures
In the second part, I introduced my project Versäumte Bilder (Missing Pictures). It began in 2022 with a photographic concept: portraying women scientists in the same way as the men typically found in the portrait galleries at many universities and institutes, in black and white, in formal poses, not smiling. These “walls of fame” are a striking example of how images shape our idea of who belongs in science.
The project took a decisive turn in April 2023 with Dr. Rosalind Franklin (1920–1958). Out of frustration at how her contribution to the discovery of the structure of DNA was not only largely overlooked but even exploited by colleagues who eventually received the Nobel Prize in Physiology or Medicine for groundbreaking DNA research in 1962, I used AI to create an image of her receiving the Nobel Prize. Seeing that photorealistic picture was an emotional experience, a kind of visual justice. It also made clear what such an image can do: it opens a space of possibility in our minds.
Today, Versäumte Bilder invites institutions to research their own history and to make overlooked women scientists visible again. The AI-generated portraits are only one small part of this process. The real work happens in the archives, in conversations with families and colleagues, and in the institutions themselves. All prompts are published, and every AI-generated image is labelled with [AI].
For Ottawa, this connection was particularly close: since 2025, I have been working on a portrait of Dr. Luise Herzberg (1906–1971), commissioned by the Science Counsellor at the Embassy of the Federal Republic of Germany in Ottawa. The day after the workshop, two framed portraits of Luise Herzberg were presented at the National Research Council Canada.
Trying It Out: AI Image Generation
The final part of the workshop was hands-on. I began with an introduction to bias in AI image generation. AI image generators reproduce existing patterns in their training data. If we do not question what they produce, they will reinforce exactly the visual clichés we are already surrounded with. Used consciously, however, they can also help us imagine what has never been photographed: the women who were in the lab but not in the picture.
AI can therefore be a valuable source of inspiration. The challenge is that these images are compelling and fascinating, yet they do not provide scientifically accurate representations. This is where responsibility comes in. Transparency is essential: labelling AI-generated content, publishing prompts and being clear about the purpose of an image are not formalities. They are what makes it possible to use these tools creatively without losing trust.
In the concluding exercise, participants visualized a scientific press release using their own AI models. Example prompts and images were presented to, and discussed with, the others.
What Remains
At the end, the workshop confirmed that we often consume images without deeper reflection, which increases the possibility of getting influenced in unconscious ways – a concern that applies not only to the underrepresentation of women in science but, more broadly, to how we perceive social injustice. However, people are ready to look more closely at images when they are given the time and the tools to do so. Reflecting on images is not about rejecting them, but about choosing and creating them more consciously.
My thanks go to the German Embassy in Ottawa and the Centre for International Policy Studies (CIPS) at the University of Ottawa for hosting the workshop, especially Prof. Jörg Esleben and Jan Völkel, and to all participants who made this workshop a lively and vivid experience with their comments, questions and creative artwork.
Gesine Born is a photographer and science communicator based in Berlin and the founder of the Bilderinstitut Berlin (www.bilderinstitut.de). She studied communication design with a focus on photography at HAW. She held a workshop at CIPS on October 6th, 2026 titled Reflecting on Images in Science – Using AI Responsibly and Creatively.
