AI = A + I: Three Questions for the Photographer's Identity

Before you read on, take a moment with the four images accompanying this article. Three of them were generated entirely by AI. One was made by a photographer, with light, lens and sensor. Can you tell which is which? Hold on to your guess. I will come back to it.

Three of the four are AI-generated. The answer is at the end.

The opening months of 2026 have been a remarkable period for anyone who cares about where photography is heading.

Around CP+ 2026, held in Yokohama from late February to early March, executives from Canon, Nikon, Sony, Fujifilm, OM System, Panasonic, and Sigma broadly drew the same line: generative AI that fabricates image content does not belong inside the camera body, even as assistive AI for autofocus, noise reduction, and authenticity tools is welcomed.

On 2 March, the United States Supreme Court declined to hear Stephen Thaler's appeal in Thaler v. Perlmutter, leaving in place the lower court ruling and the U.S. Copyright Office's position that copyright protection requires human authorship.

Then on 12 April, Boris Eldagsen and Miles Astray published a sharp public critique of Joan Fontcuberta's new book Immagini Latenti, rejecting his proposal to treat AI images as "algorithmic photography" or "second-generation photography".

Three different fronts, all circling the same uncomfortable question: what is a photograph now, and who counts as its author?

I want to offer a small reframing I have been using in my talks. When we hear "AI", we naturally hear "Artificial Intelligence". In our context as photographers, we should also hear AI as A + I. The "I" is Identity. My identity as a maker. My identity as a witness. My identity as someone accountable for what I put into the world.

That reframing brings me to the three As the AASS has been working with: Authenticity, Authorship and Accountability. Each one is best read not as a rule to follow, but as a question about identity.

Authenticity. Who am I in relation to truth?

A photograph, in the older sense, is light bouncing off a real thing and striking a sensor or a strip of film. An AI image is a calculation drawn from patterns inside a model. The two can look identical on a screen. They are not the same kind of object. The camera makers at CP+ understood this clearly and chose, deliberately, to keep generative image-making out of their hardware while continuing to invest in AI that assists the act of capture. They are staking the value of the photograph on the integrity of its making, not on how the final pixel looks. As photographers, we face the same question each time we press the shutter. Am I a witness, a documentarian, an interpreter, or a storyteller working in fiction? There is honourable room for all of these, but only if we are honest about which one we are doing today.

Authorship. Who am I in relation to creation?

Thaler's particular case has now reached the end of the road in the United States. The courts and the Copyright Office have been consistent that fully machine-generated work lacks the human authorship copyright requires, even as the line for AI-assisted work continues to be tested. That is a legal answer to what is, for us, an ethical question. What part of the image genuinely came from me? Did I see it, frame it, wait for it, decide on it? Or did I describe it to a system and accept what came back? Eldagsen prefers the word "promptography" for the latter, and I think he is right that we need a clear term for it. Not because prompt-based image-making is shameful, but because folding it into the word "photography" weakens both crafts.

Accountability. Who am I in relation to responsibility?

This is the one we tend to skip. Once an image leaves our hands, it has consequences. Subjects can be misrepresented. Viewers misled. Competitions skewed. Clients deceived. Public trust in our craft can quietly drain away. The Fontcuberta debate, at heart, is a debate about accountability. If we adopt a vocabulary that softly absorbs AI imagery into the family of photography, we make it easier for synthetic images to inherit the documentary authority that real photographs spent two centuries earning. That is not a small thing. It matters for journalism, for evidence, for memory, and for the standing of every honest photographer in the room.

Now, the four images. Image B (the architectural study in black and white), image C (the studio portrait), and image D (the aerial of the river winding through the green valley) are all entirely AI-generated. Only image A (the motion-blurred autumn landscape) is a photograph. If you guessed correctly across the board, well done. If you did not, you are in good company. Even experienced judges and professional photographers struggle with this test today, and that is precisely the point. The image on the page is no longer enough on its own.

That is what the three As are for. Before I share, submit, or sell an image, can I say honestly what it is, who made it, and what I owe to the people who will see it? If yes, I have earned the right to call it my work. If not, the question is worth sitting with a little longer.

We are in a period where the technology is moving faster than the language we use to describe it. My hope for the year ahead is that APS members keep talking to each other about these questions, in salons, on judging panels, in print, and over coffee. The "I" in AI is, in the end, ours.