With the recent deluge of AI image generators upon the market, and the continued expansion of AI assisted image capture/processing, there has been a lot of understandable discussion about what impact these new technologies will have on Photography itself. These discussions range from “AI will destroy photography” and “AI image compositors cannot call themselves photographers” to a general excitement or curiosity (tinged, perhaps, with a bit of reservation). What is generally missing in these discussions is a clear definition of what is or is not photography, a benchmark with metrics for many that would appear to vary as widely as positions on image generation by artificial intelligence.

Much of the anger towards AI assisted imaging comes from the position of a perceived threat to what humans can accomplish alone, i.e. landscape photographers that see an impossible to beat challenger, stock photographers unable to compete with the a la carteness of limitless prompts, etc. These are of course rational fears for photographers that work in this way, and certainly, AI imaging will disrupt (and has already disrupted) those industries and more–though, even photographers that reject AI image generators gladly embrace AI image correction in their own work, whether or not they recognize it as such. What’s amusing about this fear, however, is that photography itself was viewed in this way for much of its early history, by painters and other traditional artists. In fact, it is generally thought that Joseph Nicéphore Niépce’s (largely considered the inventor of photography) original motivations for creating a fixed image, resulting from the use of a camera obscura, was to navigate around the technical skill needed for traditional lithography. His invention, coined heliography (photography was named later), introduced a process that a mere technician could use to produce a result far superior to those made by even the most skilled lithographers and illustrators. Sound familiar?

What these fearers of “replacisim” fail to acknowledge is the extreme advantages they have over a complete novice in using these technologies, were they only to embrace and actually use them. To put it another way, an established landscape photographer would know better the compositional requirements, sales requirements and platforms available for AI landscape photography than someone just starting out, and would merely need to identify ways to integrate the new technology into their practice, if income is the primary concern.

Among those who fear artificial intelligence’s place in modern imaging are the so-called “photography purists”, who argue that any process utilizing AI generated images cannot be called photography. What this point of view fails to understand is a traditional answer for “what is photography?”.

The processes used in producing a photographic image have varied widely since the term “photograph” was used to describe the resulting image. So much so, in fact, that a photographer from the 1850’s, using a copper plate applied with a thin coat of silver, sensitized with iodine, and exposed to mercury vapor in order to produce an image, would likely not only not understand the “magic” of modern Digital photography, but wouldn’t even see it as photography. Some photographers who worked with film during the introduction of digital cameras didn’t see digital photography as “photography” either. After all, photography, as it was originally coined, is the fixation of an image onto a physical substrate brought about by a chemical reaction to the exposure of light, while digital photography is nothing more than binary computer data. Yes, modern digital cameras, especially interchangeable lens cameras, do emulate older film cameras in how they function, but this is not exactly a requirement for digital image capture and the longer we move away from the film era, the less and less these cameras indeed share with their film predecessors.

Even with modern photographers accepting these differences, there is resistance to allowing AI imaging to be called “photography”, and generally, I would agree. These types of images, like all digital images (which require digital processing and all of the obfuscation and interpolation that allows), are more “photographic” than “photography”. Perhaps this is the description we need (and should have been already applying) moving forward, in an era that is in actuality beyond photography. Authorship is another issue altogether, and current copyright scholarship leans toward not recognizing fully AI generated photographic images as being able to be protected by copyright registration in the way that other photographic images are. I think that this is just a matter of interface, however, and perhaps application (see my image at the top of the article). Currently, AI image generation requires a set of word based prompts from a human user, but as that process becomes more collaborative (and integrative of the human user’s unique creativity), this will likely change. The question is, will photographers?