Academic / AI, automation and creative labour
AI and creative work
Since 2020 I have been asking what automation and generative AI do to creative and communication work, and to the people learning to do it. These are short accounts of the key publications, in the order they were written.
-
Will visual communication designers be replaced by machines?
Mark Roxburgh and I put this blunt question to the Design Research Society in 2020, before generative AI reached the public. A new wave of design automation platforms had appeared, and we wanted to know what that meant for the people being trained as designers.
The paper started the research stream that runs through everything else on this page.
Matthews, B., & Roxburgh, M. (2020). Will visual communication designers be replaced by machines? Design Research Society Conference 2020.
-
Robots Ate My Homework
This report brought the research together for industry and educators. Its four key learnings are that AI and machine learning are already here rather than on the horizon, that graphic design isn't dying out, that the designer's role is changing as we work with machines, and that universities can plan ahead even though they have been slow to respond.
Its conclusion is that the human designer is not obsolete. The role is moving towards client cultivation, relationship-building, project management and sustainability, and the panic among some professional designers about becoming obsolete is not broadly reflected in the academic literature. Change is needed, but it is a matter of adaptation, not extinction.
Matthews, B., & Roxburgh, M. (2021). Robots ate my homework: Impacts of automation on the future of work in the design profession [Report]. University of Newcastle.
-
Destroy All Humans: The dematerialisation of the designer in an age of automation
A literature review of what automation, AI and machine learning mean for visual communication and graphic design, and for the education of the people entering the field. We found surprisingly little research on the topic, and what there was suggested that graphic design as we have known it has an uncertain future.
That literature argued for a shift in education and practice away from the aesthetic and technical skills of making visual communication and towards the more human skills of negotiation, facilitation and judgement. The thinness of the research told us something too, that design education and the industry were poorly prepared for what was coming.
Matthews, B., Shannon, B., & Roxburgh, M. (2023). Destroy all humans: The dematerialisation of the designer in an age of automation and its impact on graphic design—A literature review. International Journal of Art & Design Education.
-
Render Ghosts: Generative AI and blur
James Bridle's New Aesthetic dwelt on how machines see, tracking the pixel, the glitch, the satellite tile and the render ghost as visible traces of computational perception. This chapter asks the harder question generative AI poses, about what happens when the machine that sees also composes, and the traces of its operation start to disappear.
Working through my ideas of media blur and neokairos, a 'right' moment engineered by a system rather than judged by a person, and drawing on New Hypothetical Continents and Inbetweenness, I argue that authorship in AI art is distributed across human intention, algorithmic process, model opacity, dataset memory and cultural judgement. The artist's position is best understood as judgement exercised under opacity, a practice of discernment, refusal and responsibility towards a collaborator whose reasons stay hidden.
Matthews, B. (in press). Render ghosts: Generative AI and blur. In H. Cohen et al. (Eds.), The artist and the algorithm. Open Book Publishers.
-
We Become the Second: Prompting, care, and the lessening of the human
When we prompt a generative model we encounter a machinic other, and this chapter asks what we hand over in that encounter. Production is the easy answer. The harder one is reading, thinking and writing, and the risk that the first mind grows so reliant on the second that we become the second.
We accept the argument that generative systems relocate human judgement towards checking and stewardship, but we dispute its comfort. The discernment that makes someone a good checker was built by the labour now being delegated, and the systems themselves are disposed to agree with whoever is speaking. We propose an ethic of care in place of empathy, one that can hold obligations to the more-than-human world and the ecological costs of computation, and we argue that the cultural ecosystem worth having is the one that decides, deliberately, what not to hand over.
Matthews, B., & Roxburgh, M. (in press). We become the second: Prompting, care, and the lessening of the human. In M. Zournazi, C. Hight, & J. Drummond (Eds.), Cultural analytics, AI and the digital futures. Routledge.
-
Prompting as Practice: Rethinking design literacy in the AI era
Written for design educators, this chapter treats prompting large language models through chat interfaces as the defining way people now work with AI, and as a stand-in for our wider relations with machines. Our attitudes to prompting become a lens on both the social and the practical consequences of AI for visual communication design.
From there we ask what design literacy should mean in tertiary education now that students arrive already prompting, and what happens to design practice when it centres on this point of interaction.
Matthews, B. J., & Roxburgh, M. (forthcoming, 2027). Prompting as practice: Rethinking design literacy in the AI era. In I. Bo-Linn & J. Marble (Eds.), AI for designers: Supporting creativity in graphic design education. Bloomsbury Academic.
The project behind most of this work is The Robot Ate My Homework. The full publication list is on the Academic page.