In a similar vein as @averyseasonart, I entered the prompt “a black and white photo of a white man on the computer using Midjourney to generate black and white portrait photography” into Midjourney and upscaled one of the results. (edit Rhea Nayyar/Hyperallergic using Midjourney)

Jos Avery was surprised when his portraiture account amassed nearly 30,000 followers in just five months. The self-described photographer primarily posted heavily retouched black-and-white portraits accompanied by fictional stories about the subjects to @averyseasonart. But Avery recently came clean and told the world that his “photos” were actually generated by Midjourney, a text prompt-based artificial intelligence image-generation program.

Avery spoke his truth in Ars Technica last week, telling the publication that he originally wanted “to fool people to showcase AI and then write an article about it.” He did not immediately respond to Hyperallergic’s request for comment.

To develop the portrait images, Avery would submit hundreds of carefully worded prompts for image generation on Midjourney, composite the best iterations, and edit the layered pictures using Adobe LightRoom and Photoshop. On average, it took Avery about 85 generated images (~21 prompts, as one prompt yields four images) to find one that was usable and worth editing.

“It takes an enormous amount of effort to take AI-generated elements and create something that looks like it was taken by a human photographer,” he told Ars Technica. “The creative process is still very much in the hands of the artist or photographer, not the computer.”

An example of the “photographic” content from the @averyseasonart grid with no mention of AI generation (all screenshots Rhea Nayyar/Hyperallergic)

Yet, Avery deliberately identified himself as the “photographer” behind his portraiture, only revealing that his images were generative composites in his Instagram comments section three weeks ago before recently adjusting his account bio to include the term “AI.” According to Ars Technica, Avery maintained that he used “a Nikon D810 with 24-70mm lens” to capture his “subjects,” further backing himself into a corner as each post received dozens of praising comments until his admission of guilt.

Avery still posts generative portraiture on his Instagram, now hinting at the process in the captions but disclosing that the images are from AI in the comments section.

Unsurprisingly, this revelation is particularly polarizing as the conversation regarding the blurry ethics of AI use in the art sphere continues to unfold. Despite the partial commenting limitations on Avery’s account after the news broke, some Instagram users have decried his craft as “disingenuous,” “intentionally misleading,” and “a fraud,” while others remained impressed with Avery’s prompt-engineering and the lifelike results that followed.

To develop the portrait images, Avery submitted hundreds of carefully worded prompts for image generation on Midjourney.

In both his interview and his comments section, Avery vehemently defends his “work,” noting that the people who have issues with AI’s integration into art echo the same concerns as those who were skeptical or dismissive of photography when it was first invented. He also mentions that he started out as an AI skeptic himself.

“I honestly did not like AI much when I started the journey but have fallen in love with the limitless creativity,” Avery wrote in response to a critical comment. “I also like that it doesn’t create perfect images the exact way I want them. It means I have to work for it, re-generating, editing, drawing, compositing. I enjoy the collaboration.”

Spanish photographer Silvia Catalán, who criticized Avery’s content, told Hyperallergic that generative images in place of real photography “produce sadness, anger, and despair.”

A comment exchange between @averyseasonart and critical Instagram users from Avery’s recent Instagram post

“It seems good to use AI as a resource, for example, photographers have always used resources, but to use it as a photo, to make others believe that you are a photographer, or that you are a better photographer, seems very wrong to me,” Catalán continued.

Bavaria-based photographer Dirk Kultus didn’t buy Avery’s defense about working concurrently with AI to create something new. “Even if AI-generated images are modified, photoshopped or edited – they are still generated by the AI and not created by the user,” he said to Hyperallergic.

“I think it also shows how easily the AI itself fools you by giving you the idea and sense of creating something when using it,” Kultus explained. “But you don’t. You just feed the algorithms which then calculate a result based on pure randomness and probabilities on a given database, created by others. It just modifies and manipulates existing results — texts, images, whatever — and generates a variation or combination. A process almost completely dependent on the algorithms of the AI and you as a user have only a very limited influence on.”

The Latest

Required Reading

This week, the world’s lightest paint, Pakistan’s feminist movement, World Puppy Day, and were some of Vermeer’s paintings created by his daughter?

Rhea Nayyar

Rhea Nayyar (she/her) is a New York-based teaching artist who is passionate about elevating minority perspectives within the academic and editorial spheres of the art world. Rhea received her BFA in Visual...

Leave a comment

Your email address will not be published. Required fields are marked *