Ai portrayal project –/imagine
Portraits by Amit Lennon and an Artificial Intelligence engine.
This project explores the idea of imagination as used in Ai imagery and highlights the system’s inbuilt bias.
The imagery generated by the Ai makes assumptions and corrections based on its data set – which is retouched, enhanced, mostly western, and weighted towards repeated cliches. It struggles with ethnicity, and gravitates to rigid formats, such as actors’ photos.
When presented next to a studio portrait of a real person, with unique features, it exposes the possible future idea of beauty, reality and visual truth that an Ai may /imagine. These portrait pairings are a studio portrait of a real person, next to an Ai generated image. The starting point for the Ai was the real person’s biometric data, the three key face recognition data points –
Distance between the eyes
Distance from the eyes to the tip of the nose
Width of the mouth.
The rest of the image is generated based on text prompts, which include age, profession, location, and ethnic descent. The image created was based on what the AI thinks those words mean visually. Portrayal – ‘The way that someone or something is described or represented in an artistic work’
Find out more: