IN A parallel universe, these faces would belong to the well known stars of Hollywood.
But in our world, the eerily familiar faces were spawned by artificial intelligence.
The current boom in AI systems has shown machine learning algorithms are capable of astounding feats. From being more capable than humans at detecting certain types of cancer to becoming the world's best player of an intuitive ancient Chinese board game, machines are eclipsing human potential before our eyes.
AI systems have also proven quite adept at not only recognising faces, but reproducing their own. The latest example comes from US chipmaker Nvidia, which has published a paper showing how its AI system can create photorealistic pictures of fake celebrities using self correcting networks.
The researchers used what's known as a generative adversarial network, or GAN, to produce the lifelike pictures after feeding the machine a bunch of different real-life celebrity photos.
GAN systems are made up of two networks, one that produces the image and a separate discriminator network that determines how real the image looks, effectively prompting the generator system to improve its output.
The video below shows the process Nvidia's GAN went through to develop the fake celebrity photos, and also shows how it can be applied to objects and scenery.
This isn't an entirely new trick but researchers say these are the most convincing and detailed pictures of their type ever made, according to The Verge.
"While the quality of our results is generally high compared to earlier work on GANs, and the training is stable in large resolutions, there is a long way to true photorealism," the researchers wrote.
"That said, we feel that convincing realism may now be within reach."
As this technology improves, there are obvious potential benefits for companies in the creative industries. For example advertising, movies and video games could be seriously aided by the use of AI image production. Soon it might not be so clear if the pretty woman selling you shampoo is actually real or not.
But on the flip side, it does raise questions about how such technology could be used in the era of fake news. Sure we can alter images in Photoshop already but the ease with which machines could be designed to do it could lead to a Photoshopped world on steroids.
Update your news preferences and get the latest news delivered to your inbox.