Ways of [Machine] Seeing: Learning experiments in computer vision

ReMap member Annie Davey is a Lecturer in Art and Education at the UCL Institute of Education. Annie describes the project Ways of [Machine] Seeing which responds to the urgent need to update methods of teaching and learning about AI. 

Taking John Berger’s 1972 book and BBC TV series Ways of Seeing as a starting point, the research positions art and design as a vital subject to explore the ways that seeing and making images are being transformed by machine vision. Recognising that computer science has an epistemic monopoly on the teaching of AI in schools, my part in this research has been to engage art and design teachers critically and creatively with AI in the classroom.

In 2022 we were funded by The Alan Turing Institute to run a series of workshops at The Photographers Gallery with PGCE Art & Design trainee teachers. These were centered around cultivating understanding of the ‘back end’ of AI image generation; the politics of data sets, hidden labour of image annotation, algorithmic bias and the intersectional discrimination it produces. We modelled embodied, collective learning activities aimed at rendering machine learning visible, and incorporated ways of feeding the machine differently according to ambiguous, critical and poetic interpretation of images. The workshops informed the co-design of an online toolkit produced with and for teachers, which took the form of a dedicated wiki to document activities and share resources to be used by teachers in the classroom.

Whilst the workshops were well received, on reflection the activities felt quite distant from the material practices of the art and design classroom.

And, whilst the decentralised ‘politics’ of the wiki were in keeping with the project aims, its format was not user friendly for busy teachers. Most of all it was very apparent to me that whilst we had promised to ‘co-design’ with teachers, this had been curtailed by the timeframe. More time and space was necessary to understand teachers experiences, interests, expertise as well as the material and ideological limitations of the post-Gove classroom. It seemed imperative to address this imbalance so as to sustain this research and have genuine impact.

In 2023 I was awarded a UCL CCM Seed Impact Grant to extend this work via two workshops with a small group of art & design teachers and the commission of a poster and website. This has enabled us to slow down and learn from teachers. The artist Dean Kenning was approached because of his practice based research with technology, diagramming, philosophy and sculpture, as well as his understanding of critical issues in art education articulated through his writing and teaching practice. We decided to turn away from screen-based activities towards drawing and diagramming to facilitate creative, critical and conceptual understanding of AI using the tools and processes with which teachers were familiar. We were particularly keen on working towards a physical poster – a classroom object that would provide an imaginative prompt, rather than an infographic explainer.