(Pocket-lint) - Fake news, deep fakes and AI-generated content. The web is rife with these things.
Now generative adversarial networks are being used to create realistic things that don't actually exist as well. There are all sorts of them too, some are incredible, some daft, all a certainly interesting.
We've collected a few for your enjoyment.
This Cat Does Not Exist
Cats. You either love them or hate them. How do you feel about a cat that doesn't even exist though?
These AI-generated cats are fairly convincing and they're not likely to cause you any allergy problems, which is a bonus.
This Pony Does Not Exist
This Pony Does Not Exist is a site that displays a colourful grid of AI-generated pony portraits created using Nvidia's StyleGAN2 architecture.
With some clever wizardry, these creatures have been created, cropped, scaled and displayed in all their glory for your viewing pleasure.
This Car Does Not Exist
Here's a site that generates vehicles that never existed. All sorts of weird and wonderful cars that you've never seen on the road and never will.
This certainly isn't the best one on our list though. Click through and refresh the page a bunch of times and you'll see some really odd wheeled machines.
These people do not exist
The people that you see in this collection of images do not exist and never have. They have been created by a generative adversarial network that was originally developed by Nvidia, then later used by Phillip Wang to create This Person Does Not Exist.
That site essentially uses artificial intelligence to randomly generate completely artificial and yet hyper-realistic portraits of people. People who have never graced the planet and yet could quite easily live next door to you.
This technology is quite disturbing when you consider what Inverse noted when interviewing the creator.
"In a society where pictures and images are the standard surrogates for “proof,” GANs....will soon allow anyone to furnish “proof” that any imaginable person did any imaginable thing....As an example, a nefarious actor could spread a GANs-generated video or image depicting a bogus event to incite riots, protests, or other potentially violent reactions online."
Scary, but it doesn't take away from how impressive this technology is.
This Meme Does Not Exist
Everyone loves a good meme. So why not apply artificial intelligence to create some new ones. That's just what ImgFlip has done.
This meme does not exist is the result and the images generated by the system are, well, something special.
This City Does Not Exist
This City Does Not Exist is an impressive website showing various aerial images of Cities that don't actually exist. And yet they look like something you'd find by scrolling through Google Maps or peering out of an aeroplane window. Land and sea collide with manmade structures, all created by artificial intelligence.
These Eyes Does Not Exist
This one might seem unusual when compared with the others, but it's actually quite interesting.
The author/creator of This Eye Does Not Exist explains it nicely:
"Imagine that you are a makeup artist of some sort and you need to conceptualize some new eye makeup brand. But you have a problem, you are out of ideas. Generative adversarial networks allow you to create stuff that doesn't exist, lets say, photographs of eyes with makeup. This algorithm essentially allows you to "dream up" new stuff..."
It's nice to think that generative adversarial networks could be used for good. As well as creating fake nonsense out of the ether.
These Beetles Do Not Exist
This is an interesting project which uses machine learning to generate illustrations of beetles that don't actually exist. This was created using machine learning to analyse thousands of illustrations already in the public domain.
Inspired by zoology, botany and archaeology imagery, This Beetle Does Not Exist is certainly fascinating and stands apart from the others on this list that exist as somewhat convincing real-world images.
These Paintings Do Not Exist
This interesting addition to our list takes something that doesn't exist and turns it into a physical thing that does.
These are works of art created using Google's BigGAN (Generative Adversarial Networks) then printed and painted for your enjoyment by Ganvas Studio.
The idea is to use Joel Simon's Ganbreeder system to create your own artwork, then get Ganvas Studio to print it.
Ganbreeder lets you merge image and use AI to create something entirely unique - and a bit weird too.
These Portrait Paintings Do Not Exist
Art Breeder is an oddity. A site which creates fake artwork with simple tweaks to settings to change aspects of it. The results are weird, wonderful and certainly unique.
These are examples of portraits that were created by this technology but otherwise would not exist.
These Landscape Paintings Do Not Exist
Like the portraits, these painted scenes of colourful and beautiful landscapes do not actually exist but instead have been created using GANS and human interaction.
This is not a Rembrandt
This might look like a painting by master painter Rembrandt, but it's actually been crafted by technology.
400 years have passed since Rembrandt's last painting, yet art historians and technicians have managed to use a mix of data and facial recognition technology to analyse 346 different paintings by Rembrandt.
That data was then used to identify the typical geometric patterns used by the painter with a focus on certain human facial features. A 3D printer was then used to paint 13 layers of UV-ink to create a realistic painting that would form The Next Rembrandt.
This Vessel Does Not Exist
Derek Philip Au, creator of This Vessel Does Not Exist explains the logic of this one nicely:
"The history of ceramics is one of imitation and reproduction.
The apprentice obtains mastery of the craft through repetition, gradually improving their technique. Guided by a lifetime of working in the craft, the master examines each piece made by the student and throws away those deemed unsuitable.
The forger creates replicas and tests them in the marketplace. The connoisseur, informed by decades of experience dealing with antiques, judges the replicas. Those that are mistaken as authentic are sold, and the forger goes on to create even more convincing copies."
Here GANs have been used to create fake vessels that are then generated, judged and improved upon. The result is a seemingly endless collection of vases and vessels all created by artificial intelligence.