Humans of Uncanny Valley is a collaboration between interdisciplinary artist Darren Tynan and the solo electronic project of Sage J Harlow.

Masahiro Mori’s uncanny valley hypothesis states that affective responses such as unease and eeriness can arise when replicas and representations of humans appear or behave in non-human ways.

The artists explore the concept of the uncanny through computer-generated digital images, video animations, and soundscapes. Tynan curated and animated digital images of human-like faces created using Nvidia’s StyleGAN2, a Generative Adversarial Network or GAN. This form of generative image modelling can produce highly realistic synthetic images of human faces that are difficult to distinguish from authentic ones. Despite this, some GAN images are characterised by perceptual mismatch phenomena that contribute to the uncanny valley effect, exhibited in this practice-based work.

The audio component of this exhibition was created using the raw material of freely available field recordings sourced from The Internet Archive. The field recordings were time-stretched and EQ’d resulting in sounds that have a similar sense of the uncanny—recognisable, but not quite right.

Humans of Uncanny Valley aims to generate discussions about the uncanny valley effect through collaborative art practice, to showcase perceptual mismatch phenomena in GAN images, and to raise awareness about the capabilities of publicly accessible machine learning technology.

Using Format

Shopping Cart

    Your cart is currently empty

    Add an item to your cart to make a purchase.

Order Subtotal
$0.00 AUD