Winter 2025

Ultra process

Tool, agent or end of days? John L. Walters asks designers and image-makers how and why they use visual AI and what this might mean for creativity [EXTRACT]

Right at the end of a conversation about text-to-image programs with photographer Rankin, he asked what I thought. This was one of dozens of interviews I conducted for this feature, and there was no satisfactory conclusion in sight. Several exceptionally gifted and clever people have helped me put this together, but there will be no neat summary about visual AI and how it might change graphic culture. Is what we see at the moment merely the equivalent of 1990s web design, early cinema or Fox Talbot? There is a long history of over-optimistic predictions, from decision-making machines to self-driving cars.

Yuval Noah Harari, philosopher, historian and author of Nexus says that AI is the first technology in history that is not a ‘tool’ but an ‘agent’: it can invent new things. The atom bomb could not invent the hydrogen bomb, for example. Andrew Smith, author of ‘coding odyssey’ Devil in the Stack, thinks that computers are nowhere near having the learning intelligence of bees or cuttlefish. His fear is that as computers fail to become more like us we might meet them in the middle – humans will become dumber to make use of AI applications.

For the purposes of this article I am concentrating on visual AI. Designers are approaching text-to-image software with a heady mix of excitement, curiosity, disgust, unease, pragmatism, disquiet and awe – and that is before we even start discussing the negative side, such as the sustainability issues addressed in J. P. Hartnett’s ‘Baked in’ (pp.70-71).

In Marian Bantjes’ ‘Artificial idiot’ (Eye 105), she and Design Army’s Pum Lefebure likened Midjourney’s progress to that of a child. To extend the metaphor, is AI in danger of evolving into an unsocialised ‘kidult’, consuming vast amounts of ultra-processed food, unable to engage with all but the most dedicated carer and making what Max Read, in a widely shared New York magazine article (‘Drowning in Slop’, Life in Pixels, 25 September 2024), termed ‘slop’?

[…]

Read the full version in Eye no. 107 vol. 27, 2025

Top. Image from @wonderlandbyhey by Verònica Fuerte. Right. Right. Image from The Cottingley Fairies Reimagined, for which photographer Rankin worked with model Heidi Klum and FTP Digital to make images that mixed AI with genuine photographs.

Historical Surrealism. This AI-generated image is from We Are At War, by artist Phillip Toledano. ‘The fact that I was fabricating a possible fabrication was strangely interesting,’ says Toledano.

Eye is the world’s most beautiful and collectable graphic design journal, published for professional designers, students and anyone interested in critical, informed writing about graphic design and visual culture. It is available from all good design bookshops and online at the Eye shop, where you can buy subscriptions and single issues.