TESTING AN EXPERIMENTAL TEXT-TO-IMAGE MODEL


The same prompt. One year apart.


Welcome to The MoSAIC’s first virtual exhibition featuring the experimental text-to-image artificial intelligence model DALL-E v2.exp by OpenAI. This exhibition is an exciting opportunity for young people to explore the capabilities of AI and how it can be utilised in art.

The exhibition is divided into different sections, each highlighting an aspect of the model that has changed since its original launch in 2022. The paired images displayed in each section are a great way to observe and compare the original DALL-E v2.0 on the left with the updated experimental model DALL-E v2.exp on the right.

The pairs of images were created using the same text prompt, created by various artists, students, and even OpenAI's ChatGPT chatbot. As you navigate through the different sections, pay attention to the themes presented and use them to guide your critique. But also, take notice of the overall quality of the imagery and the subtle differences between the two models.

The rapid evolution of the DALL-E model in just one year is remarkable. Through this exhibition, we invite young people to think about the potential of AI in art, its ethical implications and its impact on the art world.

Perhaps the first thing you will notice about the experimental model is the dramatic improvement in image resolution.

How does this change your perception of the images?

“A computer would deserve to be called intelligent if it could deceive a human into believing it was human.”
Alan Turing

Next
Next

Lighting and Shadow