AI-Generated Content 101 — Let’s speak about the apps?
It was nearly seven years ago that the first face swapping apps appeared on the market; and since then millions of us have spent millions of hours becoming someone else.
Face Swap Live was one of the early releases and quickly became “the app that has launched a thousand YouTube videos, Imgur posts and nightmare memes”, according to The Washington Post.
Face swapping was a fascination long before this release but convincing swaps were the result of hours of Photoshop fiddling whilst everything else was funny, terrifying and totally unrealistic in equal measures.
As we chuckled at ourselves with bunny ears, a team of friends in Kyiv set about to make more realistic or deepfake technology accessible. Reface was the result of their endeavors and clocked 100 million installations in just one year. Snoop Dogg, Justin Bieber and Miley Cyrus are amongst the celebrity Refacers.
The team’s ambitions were stellar. Roman Mogylnyi, CEO and co-founder of RefaceAI, told TechCrunch: “In our dreams and in our vision we see the app as a personalization platform where people will be able to live different lives during their one lifetime. So everyone can be anyone.” You can make art talk with Wav2lip and animate your favorite gaming characters with lip syncing app Wombo.
Apps like ArtBreeder take Machine Learning (GANs) into the purely creative sphere. It promises you can “change anything about an image just by modifying its “genes” — whether portraits of people or landscapes.
This is about creating the fantastical with the help of AI.
With the charge to make the technology accessible, the realism of creations has had to be sacrificed. For the moment, the other worldliness — or just plain naffness — of some of the creations is part of their appeal. But looming large is the possibility of bringing hyperreality to these kinds of playful apps. We are already at the point where accurate voices can be synthesized from just a few seconds of recording — Lyrebird is one startup doing this — or believable entirely synthetic voices created. You can create an animated hyperreal avatar from just one image.
There is a nefarious flipside to the frivolous, however. Just recently, Vice detailed how FaceMagic, an app on Google Play and The App Store, was advertising on Porn sites offering “deepfake porn in a sec”.
Metaphysic’s Henry Adjer detailed the extent of deepface image abuse — a creeping darkness — in the BBC documentary.
Those whose images have been taken without consent remain in a legal no man’s land. This is new territory and the law has yet to venture into it sufficiently.
Hyperreal synthetic media — user generated or otherwise — is exciting and mind-expanding, despite this. We will be able to sharpen our ability to suspend disbelief and more deeply immerse ourselves in imaginary worlds. We will be able to import our hyperreal likenesses into apps accessible on our phones. This will change the landscape for app developers and raise the bar in terms of engagement and creativity.
As Reface cofounder Dima Shvets described it to journalist John Koetsier: “We want every human to have [his/her] own Hollywood in a mobile form.”
Join the Every AnyOne’s Community
Katie Scott is a Journalist with more than twenty years of experience writing on everything from tech startups to travel. Former News Editor of Wired.co.uk, now based in Sussex after a spell in Hong Kong.
