The images in a newspaper report about plans for a “55,000 acre utopia dreamed by Silicon Valley elites” caught my eye at the weekend. The ‘utopia’ is called ‘California Forever‘ and its promotional images are striking. They’re a kind of 1930s graphic-cum-landscape, pastel colours, little shading. Bits remind me of iPad painting software, others of English interwar driving guides. But mostly I thought – they must be generated by AI, a generative AI like Midjourney, Dall-E or Stable Diffusion.
I think there are two reasons for this. One is the visual content. There are weird shapes in every image: look at the green things on the roof of the building in the image above. There’s a canoeing scene in which the figures in the kayak are splodges and the daisies next to them oddly large. In fact, in each of the images on the California Forever website I spotted some strange shape which doesn’t work as a representation of even an imaginary urban space.
Also, there’ is a slightly odd ‘utopia’ vibe to the images which is not dissimilar to something Roland Meyer has identified as typical of Midjourney’s generated images. As Roland has demonstrated, Midjourney seems to like an early- to mid-20th century US public visual culture style (with apologies to any art historians out there wondering what on earth I’m talking about). It’s kind of heroic public art/messaging, but more suburban and, as Roland also showed with his image of a ‘family barbecue in a redwood valley’, with very traditional social content: white straight families, in Roland’s example. Roland also shows how Midjourney’s images can become really strange if expanded. These images of California Forever are much less surreal (overtly anyway); it’s all pastel colours and sunshine and sunsets (hence my reference to a certain kind of English interwar aesthetic). But the very conventional and normative social content remains, albeit specifically West-Coast-ised. It’s all very trad, men installing solar panels, women chatting in the town square. And no mention, textual or visual, of the Patwin tribe who lived on the land when the Spanish arrived.
Why use this visual style? Well, one thought is that there is one other recent proposal for a very large-scale urban redevelopment project which also promises to bring a green utopia to a dry land: Neom’s The Line. When I’ve given various talks about the digital visualisation of new urban development projects, it surprises me how many people mention The Line – and not in a good way (see too the comments made about its YouTube promotional video). The many criticisms of this sort of urban utopia vision might have fed into existing widespread scepticism about computer-generated images of new urban redevelopment projects that go for the ‘photo-real’ sort of look that the Neom promotional materials adopt (although always saturated with a golden glow – there’s something to say here about the aesthetic that Unreal Engine has enabled, I think, too). They are perhaps beginning to look more absurd and even offensive than simply a bit of typical marketing to be cynical about. So maybe the developers of ‘California Forever’ want to distance themselves from that critique and that visual style. (I recall that Google’s images for its ill-fated Sidewalk project in Toronto also used what looked like pen-and-watercolour sketches as well as CGIs in their publicity materials, perhaps for the same reason.)
There are surely other reasons too, though, for the use of these visuals. Not least, that using AI to generate these sorts of illustrative images – if that’s what was done – is surely much cheaper to the developer than paying a visual artist to do the work.
But also, if, as Roland suggests, all you need is a vibe from your image, then AI is your tool. And since urban (re)development is increasingly working with the aesthetic and atmospheric (as Monica Degen and I argue in our book The New Urban Aesthetic), then vibe is indeed what you need. So it’s an interesting moment for the production of images of new development projects. Accuracy or vibe? But of course both are now produced digitally so in some ways it’s no choice at all.
visual/method/culture 