These Landscapes Do Not Exist—They Were All Made By AI
By Mikelle Leow, 18 Aug 2022
Subscribe to newsletter
Like us on Facebook
Image generated by Aurel Manea and featured with permission
Worthy of joining those beautiful nature wallpapers that come with your computer, these “landscape photos” have one key difference: none of them are of real places. They were, in fact, all created from a text-to-image generator à la DALL-E and Midjourney.
Curious to see how the future of landscape photography may unfold with the exciting new wave of AI art technology, Aurel Manea, a 3D artist and landscape photographer based in Bucharest, Romania, entered some prompts into a brand-new tool called Stable Diffusion and watched it do its magic.
Images generated by Aurel Manea and featured with permission
The results speak for themselves. They’re virtually indistinguishable from the high-definition photos of natural scenes that people download as wallpapers, with an added touch of surrealism.
The beauty of AI generators is that you can introduce some elements of fantasy into these so-called “photographs” without having to manually add them in—all within a few seconds.
Image generated by Aurel Manea and featured with permission
In some of the images, you’ll find lush, green grasslands overlooking waterfalls, contrasting with snowy mountains in the background. Abundant crops of trees are seen standing virile on ice.
Image generated by Aurel Manea and featured with permission
There’s one kink that Stable Diffusion developers might need to iron out while the software remains in beta mode; Manea points out that some of the reflections on the water are still a little off.
Stable Diffusion, developed by London and Los-Altos-based Stability AI, seems more adept than DALL-E at producing realistic landscapes as the tool is open-source, has a larger image base, and has no content filter, PetaPixel reports.
The app is being envisioned as a more accessible text-to-image model where “billions of people” will be “empowered… to create stunning art within seconds.” As it stands, all prompts fed to the machine and results generated from the software are public domain, Manea tells us.
The artist also took some prompts suggested by DesignTAXI and ran them on the machine. In just a couple of seconds, the tool produced a series of imaginary scenes you wouldn’t be able to find out in the wild. Here are some of them:
Check out more landscapes that the artist churned out using the AI software, and head over to Aurel Manea’s website and Instagram to discover more of his work.
Image generated by Aurel Manea and featured with permission
Image generated by Aurel Manea and featured with permission
Image generated by Aurel Manea and featured with permissionImage generated by Aurel Manea and featured with permission
Image generated by Aurel Manea and featured with permission
Image generated by Aurel Manea and featured with permission