By Xavier Lam
On October 25th, this painting was sold for $432,500.
It was auctioned off for over 45 times its estimated value of $7,000-$10,000, which may be due to the novelty of how it was produced. The painting was created using an artificial intelligence software developed by the French art collective Obvious.
Using Generative Adversarial Networks (or GANs), Obvious created a machine that was capable of producing portraits out of the 18th century. GANs work by putting two algorithms that compete and train against each other. One of them, a ‘generator’, produces images based on a training dataset. The other algorithm, a ‘discriminator’, judges whether the images produced are “real”. The two algorithms loop until the discriminator can not tell the difference between a real image and one created by the generator.
In my opinion, the fact that AI can create art poses big question. Is AI capable of creativity? What is creativity? Is it possible for AI to become as advanced as the human mind? The list goes on.