We can initiate AI with text prompts or by clicking a camera shutter. Why is the output of only one of them eligible for copyright?
I think that the sense of ownership of intellectual property derives from deep features of consciousness and personal identity that are very real but also esoteric and difficult to discuss.
Rather than being driven by explicit decisions that a photographer makes, or is assumed to make, I think that the photographer's proprietary connection to their photographic image is a reflection of a moment in shared conscious experience - an aesthetic-participatory phenomenon in which something of nature or humanity is revealed through the sensitivity and motivation of the photographer. The photographer is showing the audience what they saw, or could have seen if they were able to see as quickly as a camera can register physical photosensitivity.
The photographer succeeds when they capture a moment that is rich in visible Significance, or said another way, a moment where Significance (aesthetic-participatory 'saturation') is presented visibly, making the ephemeral moment and the significance of it potentially eternal.
What I'm trying to get as is that I think that the proprietary essence of the artist is hinted at holographically in their art. A single tiny gesture can reveal it if the audience is familiar with their body of work. In the past, that might be called a soul, but I think that AI (among other things) is pushing us to really understand what that word means in scientific or at least legal terms.
Before the digital era, a person's signature was used because the nuances of the gesture made with ink on paper were so subtle and seemingly unique that they were not easy to forge perfectly, and also because the relationship between a person and their own name has a particular significance. Graphology was quite popular in the past for that reason, even though it is now popularly held to be pseudoscientific and unreliable.
The questions that AI brings up for me relate to the ability of a generated content to carry that esoteric, gestural, holographic quality of personhood. Looking at something like a Rothko painting, there seems to be some numinous quality that comes through even in the supremely simple-seeming presentations.
In the case of an AI prompt, what seems to be happening is that while the invention of the prompt itself can carry some proprietary qualities, most of the value actually comes from the larger context of shared consciousness. At this point it seems that some people are just especially 'good at it'. In the hands of an artistic person, AI can act as a camera does for a photographer, but instead of capturing an intimate fragment of direct relationship to nature, there is an intermediate Baudrillardian/simulacra layer of data that is ultimately extracted, often unwittingly, through public and private surveillance. The "light" of the AI camera does not belong to nature but to human beings, both in the public community and in the ranks of developers and engineers who have built it, often for commercial or other politically charged purposes.
It is hard to imagine any kind of satisfactory solution to the AI/IP legal tsunami. If I had to guess, I would think that whoever is willing to spend the most money or bring the most attention to their individual case will come out ahead for the foreseeable future. Our legacy legal and political systems are not equipped to deal with the politics of human consciousness as it actually is, but I think that is what is ultimately necessary. We need a 21st century Manhattan Project on the intersection of consciousness and mental health, technology, money, and politics.