Re; On AI in response to: A Positive Technologist Identity (2/4)
Nicholas A. Ferrell
March 24, 2026
(See part 1 of my reply.)
Mr. Powiertowski writes that the point on which he would “push back gently is on the implication that AI-assisted media is somehow lesser for being assisted,” using early negative views of photography as an example. While my largely negative views of the current AI moment in the context of online writing and “content” are well-documented (including in links from A Positive Technologist Identity), I will note that I cited to Mr. Powiertowski’s “thoughtful” article as involving a “non-offensive” use of AI. My myriad critiques of common AI uses were not the central point of my original post.
In his gentle pushback, Mr. Powiertowski analogizes AI today to photography in its early days. “When photography emerged, painters dismissed it as mechanical reproduction, not real art, because the image wasn’t made by hand.” He noted, correctly, that the standing of photography as a field improved, and today it is viewed as an art-form distinct from painting and drawing (I am somewhat familiar with the context of his example thanks to my good friend Victor V. Gurbo).
I would respectfully suggest a different analogy to AI and writing than did Mr. Powiertowski. As I see it, AI is not akin to photography. There is no one “behind the camera” of AI and without going into the particulars of nineteenth and early-twentieth century critiques of photography, I am comfortable in drawing a category distinction between prompting a chatbot and using a camera to capture a moment before processing the image (or in Mr. Powiertowski’s words, which are backed up by his own aesthetic photography portfolio and Pixelfed profile: “The person behind the lens still chose what to point it at, when to press the shutter, and what story to tell”). If we are using photography as the example, I would compare and taking a photograph to using AI tools to re-touch or re-scale the human-shot photograph to prompting an LLM to generate an image and then editing the LLM image. Similarly, in the context of writing and AI, I think the issue is whether one drafts an article and uses tools like Claude or Grammarly for proof-reading and editing, or whether one prompts Claude to write the article. While I do not personally use AI in my writing or image publishing workflows, I think there are material distinctions inherent in different ways of using AI.
Discussion in the ATmosphere