An AI Painted These London Landmarks. Can You Recognise Them?

M@
By M@
An AI Painted These London Landmarks. Can You Recognise Them?
Two colourful interpretations of the tube roundel

It's the tube roundel as you've never seen it before — two interpretations of the famous symbol, as shaped by an AI algorithm.

You've probably seen the Dream app trending on social media by now. Feed some keywords into the app, choose your style, and the AI builds up a semi-abstract image based on those keywords. The results are beguiling.

For the images above, we simply typed in 'tube roundel', and out came these twisted fantasies. How it works is anyone's guess. The bot presumably samples hundreds of public images then splices, layers, curls and blends its way to a masterpiece. It's all pretty neat.

Both concepts weave in the red, blue and white of tube carriages and twist them into the circular form of a roundel. We're particularly impressed with the picture on the left, with its twisted and ambiguous mechanical elements.

We thought we'd try the app with some other London themes. See what you think.

Two skyscrapers as painted by an AI bot
A pair of skyscrapers. Here's Dream's interpretation of the Gherkin and the Shard.
Two paintings by an AI bot. On the left is a twisted version of Tower Bridge. A pink image on the right is a fantasy of Buckingham Palace
Two older buildings: Tower Bridge on the left and Buckingham Palace on the right
Typing in London Underground extrudes some fascinating results. The filter on the left offers a steampunk interpretation, while that on the right provides a more abstract take.
And finally, we couldn't resist trying "London at Christmas", which chewed out this rather charming image. We also gave it some portraiture to handle... no prizes for guessing the identity of the twisted miscreation on the right.

The Dream app from wombo.art is available from the App Store and Google Play, or you can try it in a browser. The Verge has an interesting article looking at the future of AI-generated art.

Last Updated 23 May 2022