I take the environmental concerns around AI seriously. I grew up in the Canadian Rockies and spend the vast majority of my free time seeking out adventure in remote wilderness areas. Data centers consume enormous amounts of electricity and water, and artists should not dismiss that. But it is also important to distinguish between individual critical art practice and the much larger industrial uses of AI by corporations, banks, retailers, advertisers, cloud providers, and technology companies. The burden of AI resource consumption is structurally concentrated among companies that build, train, deploy, and monetize these systems at scale. The environmental problem is not primarily caused by individual artists experimenting with AI images; it is driven by the massive commercial buildout of AI infrastructure. My occasional use of AI is not an endorsement of that system, but a way to critically examine the image culture that we are already immersed in.
Some additional thoughts on AI-generated imagery that I think are important to consider…
The negative reaction to AI-generated photography is understandable, but it is often framed too narrowly. Many critics treat AI as a betrayal of photography, when it may be more accurate to see it as a confrontation with photography’s unresolved history. Photography has never been an innocent medium. It has been tied to possession, surveillance, colonial classification, gendered looking, documentary extraction, and the authority of the photographer as the one who sees, frames, and takes. The camera has often converted people and places into images that can be owned, circulated, interpreted, and used by others.
AI-generated photography does not eliminate ethical concerns. Questions of training data, copyright, environmental impact, labor, and platform control are real and should not be dismissed. But these concerns belong to a different ethical structure than the one inherited from camera-based photography. In conventional photography, many of the medium’s problems arise from the encounter between photographer and subject: the act of pointing, directing, capturing, exposing, and transforming a real person into an image-object. AI can interrupt that structure. It allows an artist to construct photographic images without requiring a real body to submit to the camera, without turning another person’s vulnerability into visual material, and without repeating the heroic mythology of the photographer as the one who goes into the world and takes.
This does not make AI photography ethically pure. It relocates responsibility. The artist must now address the politics of the dataset, the biases embedded in visual culture, the environmental cost of computation, and the danger of synthetic images being mistaken for evidence. But that relocation matters. AI photography can neutralize some of the historical violences of photographic practice precisely because it no longer depends on the indexical capture of a subject who stood before the lens. It shifts photography from possession to construction, from extraction to speculation, from the authority of the camera to the critical use of a system.
For that reason, AI photography should not be dismissed as merely unethical or inauthentic. It should be judged by a more rigorous standard: whether the work understands the apparatus it uses, whether it is transparent about its constructed nature, whether it resists rather than repeats inherited visual stereotypes, and whether it opens a meaningful space for thinking about images after the collapse of photographic certainty.