Synthesized Technical Image #3
2026
Image (color)
1/1
5016 × 5016 pixels
Synthetic Technical Images is a series of digital images within the larger Technical Images body of work begun in 2024. The series engages with systems, emergence, and transformation, concepts long central to Reas’ work, but transposes them into the realm of machine learning. Each image passes through a sequence of computational models, moving from visual data to synthetic image, from image to language, and from language back to image. Through this process, the work explores how images are interpreted, translated, and transformed as they move between different machine representations.
A collection of Synthetic Technical Images starts with creating a custom dataset of foraged plants. This dataset is used to train a generative adversarial network (GAN). Images generated by the GAN are uploaded to a large language model (LLM) with the prompt, “Write a prompt so that you can generate this image from the prompt.” The resulting prompt is then fed back into the LLM with the instruction, “Generate an image from this prompt.” Each Synthetic Technical Image is an unedited image generated by the image model in response to the LLM-produced prompt.
Synthetic Technical Images #1–#3 use StyleGAN 2, GPT-5.6 Sol, and ChatGPT Images 2.0.