Anya Macdougall

Executive Summary

Anya Macdougall works at the intersection of analog photography and computational generation, using open-source diffusion models to create hybrid images while actively measuring AI's environmental impact. Her path to this practice was deliberately cross-disciplinary: four years of photography school in Warsaw, completed at seventeen, followed by a degree in artificial intelligence and psychology driven by curiosity about whether she could expand her creativity beyond the analog and manual into something different.

Her practice is defined by a commitment to control and transparency. Rather than using commercial platforms, she works with ComfyUI, a modular open-source tool that lets her insert her own images, adjust individual settings, and monitor GPU power usage in real time. This technical specificity is not incidental. It shapes both her aesthetic outcomes and her ethical position.

Her orchid project made AI's environmental cost an explicit subject of the work, calculating carbon emissions from local image generation and incorporating sustainability data into the exhibition itself. Her pixelized photography series used AI primarily as a coding assistant, automating frame conversion and concatenation that would otherwise make video work impractical, while keeping every aesthetic decision in her own hands.

I really like using AI for coding and for mini efficiencies in my creative workflow. Making things faster, but still not taking all my credit for it, because as a creative I still want to feel like the work is mine.

Her relationship with AI is conditional and considered. She acknowledges shared authorship honestly while claiming credit for technical mastery and clear creative intention. Her future use of AI depends on how the companies behind these tools behave.

Background and Cross-Disciplinary Drift

Anya finished four years of photography school in Warsaw at seventeen, having trained in both photography and graphic design including darkroom work. Career direction afterwards felt uncertain. The market seemed tough for artists committed to a single discipline.

In 2020 she pivoted entirely. Driven by an interest in science fiction and games, she studied artificial intelligence and psychology for three and a half years. The aim was to contrast everything she already knew, to find out whether her creativity could expand beyond the analog into something different. She had always considered herself more creative than technical.

An internship as a creative technologist in the Netherlands gave the missing thread its name. She had never heard the term before that placement. It was, in her account, exactly what she had been looking for: something that tied technology and creativity together in the prototyping, research, and design world she had not previously known existed.

The Orchid Project

Anya's orchid project began with a serendipitous find. A vintage orchid book, picked up in a secondhand shop, struck her with the beauty of the flowers as she paged through it at home. She began intuitively overlaying the prettiest orchids onto photographs she had taken. The combination came spontaneously.

What started as play became a structured investigation into AI's environmental cost. She chose to run image generation locally through ComfyUI specifically because local generation makes measurement possible.

AI has a huge carbon footprint, especially when using huge models in data centres somewhere where you don't even actually know how much footprint is being generated.

By monitoring GPU power usage and calculating emissions against her country's energy grid, she could incorporate real sustainability data into the exhibition alongside the artwork itself. The creative process was extensive: more than a hundred generations to find images that could overlay convincingly with her photography. Her technical knowledge shaped aesthetic decisions throughout. She knew the generated orchid backgrounds needed to be mostly black with very bright flowers to achieve the depth and layering she wanted.

She applied the same experimental approach to ComfyUI as she would to traditional image editing, treating the AI tool as another medium to play with rather than a shortcut to finished results. At the start, the flowers came out artificial and overworked, too much going on. She iterated around the right parameters once she found them, changing only a couple of aspects of each prompt at a time, refining toward realism.

The exhibition presented the hybrid images alongside the carbon data. She approached an AI art publication about the practices that should perhaps accompany using these tools, the kind of disclosure standards the field has not yet developed.

Pixelized Photography

Her pixelized series began with a discovery about image compression. Resizing a photograph dramatically down to sixty-four pixels and back up to two thousand produced a distinctive blurred, pixelized layer that she liked immediately. The technique gave new meaning to ordinary travel photographs. Sunrises and sunsets that felt mundane on their own became something more universal, images other people might recognise as their own memories.

Scaling the technique to video required automation. A five-second video contains thousands of frames, and manually resizing each one was not viable. AI entered here, but as a coding assistant rather than an image generator. She used ChatGPT to write scripts converting videos into frames, used Photoshop actions to apply the pixelization process across the whole folder, then asked for code to concatenate everything back into video.

The distinction matters to her. AI handled the mechanical repetition. Every aesthetic decision, including the technique itself, the images chosen, and the degree of pixelization, remained hers.

Control, Authorship, and Collaboration

Anya maintains clear boundaries around how much creative latitude she gives AI, and those boundaries are active rather than passive. How much to take from the system and how much freedom to give it is always her decision. When AI becomes overwhelming, when it spits everything out and the volume of output becomes too much to handle, her response is precision: be as specific as possible for what you are asking it to do.

Her authorship position is nuanced and honest. She would rather describe the orchid work as collaboration than claim sole credit, acknowledging that the training data underlying any generative model represents billions of other people's work. But she also claims credit for herself.

Because I experimented with the tool, I know it, and I also knew what I wanted in my head. I give credit to myself also, not just to it.

She draws a parallel to traditional artistic influence. Van Gogh was inspired by other painters and wanted to replicate them. How much credit does he take for that inspiration? Copyright, she notes, is a huge topic and one without easy resolution. Her own resolution turns on intent and mastery. Brainstorming with random prompts and accepting whatever emerges is one thing. Working with a clear idea, deep tool knowledge, and iterative refinement is another.

Platform Choice and Open-Source Values

Anya's preference for ComfyUI over commercial platforms reflects practical and ethical reasoning at once. Commercially, the tool offers modularity and flexibility through custom nodes, tangible settings, and the ability to insert her own images and observe how they influence generated outputs.

In ComfyUI I can pop my image in there, that's mine, and then manipulate it based on the settings I control.

Ethically, the platform enables the local generation and energy monitoring that made the orchid project's sustainability dimension possible. She uses Flux within ComfyUI rather than Stable Diffusion, finding it more flexible to work with experimentally. The transparency of being able to see and adjust each module aligns with her broader orientation toward understanding what she is working with rather than accepting outputs as black-box results.

Her conditional relationship with AI extends to the companies behind these tools.

If they do very unethical things behind the scenes, I might stop using it because of these ethical aspects I'm more worried about.

This is not abstract concern. It is a live consideration shaping which tools she chooses and how she uses them.

Analog Curiosity and Future Directions

Despite her technical fluency with AI tools, Anya remains fundamentally analog-curious. She can easily imagine returning to purely analog work depending on the project, and describes herself as someone who likes to experiment with very different mediums. AI is one medium among several rather than a replacement for photography.

Her bachelor's research explored whether AI-generated imagery could influence people's behaviour toward more sustainable actions, whether exposure to partially generated images might support climate awareness. This thread connects directly to her orchid project's sustainability focus and points toward the kind of meaningful application she wants to pursue. She is not interested in using AI as a brainstormer or email-writer, things everyone already does. She wants to see how much more impactful it could actually be.

What draws her to ComfyUI specifically is the quality of attention it demands. Watching settings shift outputs in real time keeps the work feeling like hers, even when AI is generating the images. The exchange is fast, gratifying, and continuously responsive to her direction. The making feels like making, even when the medium is computational.

Branching Into Physical Builds

From a follow-up conversation, May 2026. The original interview was conducted in November 2025.

The instinct to keep the making feel like hers has begun to push the practice in a new direction. Anya's work has started to evolve into physical builds, projects that combine circuit boards and electronics with the photographic and generative threads already running through her practice. AI coding has been central to this shift, letting her learn the territory faster and actually build things rather than circling them. What once felt scary has become approachable, which has changed what feels possible to attempt.

The sustainability strand has kept pace. Google recently released a paper on the energy footprint of Gemini, and she has found it a useful guide for realistically measuring the cost of her own requests, extending the carbon-aware approach she developed for the orchid project into her daily use of language models. Her general process and outlook remain similar. AI is still a medium to be controlled, measured, and questioned. The range of what she builds with it has simply grown wider.