Deepfaking human genomes: The use of generative adversarial networks for synthesizing DNA

Why scientists are deep-faking human genomes.

Researchers have taught AI how to create artificial genomes. This could help solve the dilemma of protecting genetic information and collecting enough DNA to do research.

The Generative Adversarial Networks (GANs), pit two neural networks up against each other, produce synthetic data so convincing that it could pass as real data. There are many examples on the internet — they generate pictures and videos, such as \”This city doesn’t exist\”. AIs are able to create convincing articles, blogs about food, and human faces.

Researchers from Estonia are now going deeper with deepfakes human DNA. They developed an algorithm to repeatedly create the genetic code of non-existent people.