CDCS Annual Lecture 2023

On the 13th of December 2023 we hosted our annual lecture, a public talk given by a leading researcher in the digital arts, humanities or social sciences. This year we were delighted to be joined by Mary Flanagan, a hyperdisciplinary researcher, designer and artist based at Darmouth College in the US. Mary’s paper ‘How to See What’s Missing’ presented her practice of using games, algorithms, and digital systems as tools for participation and discovery.  

In the first part of her lecture, ‘Silent Little Judgements’, Mary used her project ‘Help Me Know the Truth’ as a case study to explore how our internal biases affect categorisation in relation to facial recognitions. She discussed how she as an artist used the tools of science to explore how we see each other. This project explored biases within facial recognition software by asking participants to categorise images as, for example, angry, upset, hopeful etc. The images presented were that of the same face manipulated to appear different, and viewers were asked to contribute their own image to the collection so that it too would appear in the exhibition. Through this project, Mary’s work created an artistic lens through which to reflexively think about these systems and their effects, helping generate ideas about how we respond to and can overcome these biases.  

Part two of Mary’s presentation, ‘The Missing Women’, delved into the lack of female representation in many of the data sets used to train AI models. As a response to this Mary created Grace, her “unabashedly feminist” AI. This is an AI model trained by a woman on a body of information created by women. Mary embarked on creating Grace because she wanted to challenge the common practice of basing a research question on the available training data. She began a long journey trying to uncover bodies of work by female artists. Through this journey she encountered challenges in finding data, often stemming from poor categorisation and labelling of female artists' work. After much investigation she discovered 25,000 pieces of data on which to train Grace:AI. Mary then displayed some of the artwork created by Grace, which included depictions of Mary Shelley’s Frankenstein, and a collection of stunning cloud images named ‘Daydreaming’. 

This presentation was followed by a thought-provoking and insightful Q&A in which audience members both in person and virtually engaged in discussions around AI, feminism, and the art world.  

We would like to extend a huge thank you to Mary for her talk and to all who attended!