I do not use AI. I have never willfully prompted an LLM and hope to never have to. The only possible exceptions I can think of are interacting with customer service chatbots when no other communication option is presented to me by the company, e.g. when trying to reschedule a missed flight. When another option exists, like explaining my problems to a human, I always take it!
I have never generated an image or video or audio clip or anything else using a diffusion model or any other technology that is trained on stolen data (or even licensed data). I have used Remove.bg on several occasions since its introduction in 2018; I am not sure exactly what technology and/or training data underlies their product, but for the last few years it has given me pause and so I have avoided it. I sometimes use dadaist or experimental techniques when making music and graphics which could be considered "generative," but this is categorically different from what we mean when we say "generative AI." I have generally thought that Markov chains were boring since before LLMs existed.
I write code for a living and also for fun, but I have never used an LLM to generate code. I do not use any AI assistance when writing code and do not willfully employ any AI models to review code, though my employer has enabled a whole host of agents for various purposes including code review, and I am sometimes subjected to the output of those agents in the course of my work. I try at all times to avoid their gaze (which is usually easy because the things they output are usually nonsense). I try at all times to encourage my colleagues to think for themselves and use their brains to do their work, except when I think it might jeopardize my standing with my employer, at which point I shut up and use my own brain to do my own work.
I generally try to avoid using or consuming others' work when it has largely been generated by AI. A poster for an event that clearly came out of ChatGPT will be one reason for me to not attend that event, but will rarely be a reason to avoid attending something I wanted to attend otherwise, but it will affect my opinion of the people putting on the event. This is generally true for movies and games that have minor AI elements as well, though I give less benefit-of-the-doubt to vibe-coded websites. If your site smells like Claude I will just close the tab. I also won't listen to your generated music even if it's so random and funny.
My reasons for actively avoiding LLMs and other large-scale generative/transformative models are many. Among them are the vast environmental impacts of the rapid increase in demand for computation; the broad and guiltless theft of intellectual property required to train the models; the ongoing psychological effects I've observed in my peers who heavily use these technologies; and the general lack of quality and/or taste in their outputs. Any of these reasons alone could be worth quibbling over; together I consider them inexcusable.
I'm writing this first version of this statement on November 10, 2026. This page will be updated if any of this changes.