Terrifying Things Happen When an AI Generates Fake Faces Synced to Music

1
109
Terrifying Things Happen When an AI Generates Fake Faces Synced to Music

We are still trying to identify the best applications for neural networks, machine learning and all recent advances in artificial intelligence. Among all Practical research there are many frivolous experiments be done with results that cross the line fascinating and terrifying.

Automated image processing has emerged as a strong combination of artificial neural networks, fueled in part by decades when everyone shared photos and selfies of each other on the internet. It has resulted in huge archives of headshots that have been harvested and used to train AIs to do everything artificially aging users in new mobile apps to generate huge collections from photorealistic headshots of people who don’t really exist.

The stock photography industry will never be the same, but Mario Klingemann wondered what would happen if those same artificial neural networks that produced fake headshots were synced to music, generating the most expressive faces when a song’s beat really pounded.

Klingemann used the StyleGAN2 generative adversarial network created by Nvidia and finally released as an open source tool over a year ago. He did not do custom image training himself but instead adjusted the GAN to adjust the output results based on the sound spectrum of a given audio file, in this case the song Triggernometry through Kraftamt.

Some of Klingemann’s followers on Twitter have suggested he roll back some of the GAN’s more extreme output results, the horrors of which will only really be revealed if you carefully go through the above video frame by frame.

They make some good points, but if you look too closely at something, you will undoubtedly find something you don’t like – or in this case, something else to avoid sleeping well at night.

.

1 COMMENT

  1. What a fantastic post! This is so chock full
    of useful information, I can’t wait to dig deep
    and start utilizing the resources you have given me.

    Best regards,
    Thomassen Valenzuela

LEAVE A REPLY

Please enter your comment!
Please enter your name here