As hybrid neuro-photography and generative visual tools become deeply embedded in contemporary creative practices, a central question increasingly demands attention: can artificial intelligence be regarded as a co-author of visual work, or does authorship remain the exclusive domain of human creative agency? The difficulty lies in the fact that neural networks now perform not merely technical functions but aesthetic and structural ones, shaping the image in ways that transcend simple execution of commands. Traditional photography links authorship to human intention — to a photographer’s control over framing, lighting, narrative, and conceptual direction. In hybrid visualization, however, AI models introduce their own patterns, biases, and decision-making logic learned from vast image datasets. This raises the issue of whether the human artist remains the sole origin of the creative act when the final image is partly determined by algorithmic processes.

The degree of human control plays a decisive role. When AI assists only with minor post-processing or enhancing dynamic range, authorship remains clearly human. But the situation becomes far less straightforward when the artist provides prompts or guidelines, and the neural network generates new visual structures and textures that the human did not explicitly foresee. In cases where the system autonomously synthesizes most of the visual content from high-level input, the notion of a single human author begins to feel insufficient. The creative process becomes distributed between human intention and algorithmic transformation.

Legally, most jurisdictions continue to treat AI purely as a tool. Copyright protection is granted only when a human contribution can be demonstrated, while fully AI-generated works increasingly fall into the public domain due to the absence of a legally recognized author. Yet this legal framework is misaligned with the practical realities of hybrid visualization, where creative responsibility and decision-making are shared across human and non-human actors. The philosophical challenge is equally pressing. If authorship requires consciousness and subjective intention, AI cannot be considered a co-author. But if authorship is defined simply as the origin of unique cultural artifacts, neural networks undeniably play a formative role in shaping the final visual product.

Some researchers propose moving beyond the binary human-versus-machine opposition altogether. Instead of naming a single author, we might acknowledge a network of contributors: the human creator, the datasets and photographers whose images trained the model, the engineers who built the architecture, and the algorithmic patterns that emerge independently from any one individual or entity. In this perspective, hybrid visualization becomes a multi-authored practice in which both human and machine agencies are interwoven.

This debate is not limited to intellectual property; it touches on identity, creativity, and cultural evolution. Whether acceptance of AI as a co-author represents a threat to human artistry or, conversely, a new paradigm of collaborative creation remains unresolved. What is clear is that the boundaries of authorship are shifting, and hybrid neuro-photography is one of the key arenas where these shifts are becoming visible.

46 minutes ago, STAYM said:

As hybrid neuro-photography and generative visual tools become deeply embedded in contemporary creative practices, a central question increasingly demands attention: can artificial intelligence be regarded as a co-author of visual work, or does authorship remain the exclusive domain of human creative agency? The difficulty lies in the fact that neural networks now perform not merely technical functions but aesthetic and structural ones, shaping the image in ways that transcend simple execution of commands. Traditional photography links authorship to human intention — to a photographer’s control over framing, lighting, narrative, and conceptual direction. In hybrid visualization, however, AI models introduce their own patterns, biases, and decision-making logic learned from vast image datasets. This raises the issue of whether the human artist remains the sole origin of the creative act when the final image is partly determined by algorithmic processes.

The degree of human control plays a decisive role. When AI assists only with minor post-processing or enhancing dynamic range, authorship remains clearly human. But the situation becomes far less straightforward when the artist provides prompts or guidelines, and the neural network generates new visual structures and textures that the human did not explicitly foresee. In cases where the system autonomously synthesizes most of the visual content from high-level input, the notion of a single human author begins to feel insufficient. The creative process becomes distributed between human intention and algorithmic transformation.

Legally, most jurisdictions continue to treat AI purely as a tool. Copyright protection is granted only when a human contribution can be demonstrated, while fully AI-generated works increasingly fall into the public domain due to the absence of a legally recognized author. Yet this legal framework is misaligned with the practical realities of hybrid visualization, where creative responsibility and decision-making are shared across human and non-human actors. The philosophical challenge is equally pressing. If authorship requires consciousness and subjective intention, AI cannot be considered a co-author. But if authorship is defined simply as the origin of unique cultural artifacts, neural networks undeniably play a formative role in shaping the final visual product.

Some researchers propose moving beyond the binary human-versus-machine opposition altogether. Instead of naming a single author, we might acknowledge a network of contributors: the human creator, the datasets and photographers whose images trained the model, the engineers who built the architecture, and the algorithmic patterns that emerge independently from any one individual or entity. In this perspective, hybrid visualization becomes a multi-authored practice in which both human and machine agencies are interwoven.

This debate is not limited to intellectual property; it touches on identity, creativity, and cultural evolution. Whether acceptance of AI as a co-author represents a threat to human artistry or, conversely, a new paradigm of collaborative creation remains unresolved. What is clear is that the boundaries of authorship are shifting, and hybrid neuro-photography is one of the key arenas where these shifts are becoming visible.

Before getting into this sort of masturbation, a start would be for AI developers to recognise the intellectual property rights of the authors whose material their robots scrape off the internet. AI is by its nature an intellectual parasite. So it's a bit rich to suggest that AI should be given IP rights as if it were an author.

3 hours ago, STAYM said:

can artificial intelligence be regarded as a co-author of visual work

No

3 hours ago, STAYM said:

or does authorship remain the exclusive domain of human creative agency?

Yes

What's the problem ?

Artist's haven't made their own paint for more than a century, no one in their right mind has suggested the Rownney or Reeves or Dulux et al are credited along with the artist.
Perhaps the picture framer should also be credited, along with the person who hammered the nail into the wall to hang it on as well as the person who made their sandwiches. At least they were all human.

3 hours ago, STAYM said:

neuro-photography

The whole post appears to be an 'influencer's' attempt to influence in an artistic subject.

What does this have to do with SF Science news section, or even Science in general ?

@exchemist +1 for the point, though I thought the imagery a tad garish.

Edited December 1Dec 1 by studiot

3 minutes ago, studiot said:

@exchemist +1 for the point, though I thought the imagery a tad garish.

Fair comment. I'm afraid I find these attempts to assert and stretch the allegedly human-like attributes of AI rather irritating, so I expressed myself strongly. It seems to be part of the climate of breathless hype around the subject. I think there is far too little acknowledgement of the absolute dependence of these machines on human intellect.

And as far as creative arts are concerned, authors and artists are rightly up in arms about the taking, and subsequent monetising, of their handiwork without acknowledgement or recompense. That is a live issue today and needs to be settled before airy contemplation of whether an AI program can be itself regarded as an author or artist.

13 minutes ago, exchemist said:

Fair comment. I'm afraid I find these attempts to assert and stretch the allegedly human-like attributes of AI rather irritating, so I expressed myself strongly. It seems to be part of the climate of breathless hype around the subject. I think there is far too little acknowledgement of the absolute dependence of these machines on human intellect.

And as far as creative arts are concerned, authors and artists are rightly up in arms about the taking, and subsequent monetising, of their handiwork without acknowledgement or recompense. That is a live issue today and needs to be settled before airy contemplation of whether an AI program can be itself regarded as an author or artist.

We live in a country about to do away with juries.

Perhaps the next step is to replace judges with AI ?

This whole issue is much wider than just artists

Edited December 1Dec 1 by studiot

14 minutes ago, studiot said:

We live in a country about to do away with juries.

Perhaps the next step is to replace judges with AI ?

This whole issue is much wider than just artists

Oh sure, but the OP was about using AI in the creative arts.

Moved to ethics, because this isn’t science news

---

To echo previous points, Gen-AI is at its core plagiarism, so I’m not sure how you give it “authorship”

Do we extend authorship to spellcheck and autocorrect?

Just now, swansont said:

Moved to ethics, because this isn’t science news

---

To echo previous points, Gen-AI is at its core plagiarism, so I’m not sure how you give it “authorship”

Do we extend authorship to spellcheck and autocorrect?

In practical terms do people who upload AI generated content get rewarded for their "content"?

I recently looked through a supposed "Nov 30 episode of Kimmel" and it only became completely to me obvious in the comments that it was not genuine.

I think it was Youtube and so they may have had ads and been paid for this misleading product.

Music/art seems different but I guess those "content providers" get money out of AI plagiarisation.

I hardly click on Youtube any more unless I already know what is inside.

14 minutes ago, geordief said:

In practical terms do people who upload AI generated content get rewarded for their "content"?

I recently looked through a supposed "Nov 30 episode of Kimmel" and it only became completely to me obvious in the comments that it was not genuine.

I think it was Youtube and so they may have had ads and been paid for this misleading product.

Music/art seems different but I guess those "content providers" get money out of AI plagiarisation.

I hardly click on Youtube any more unless I already know what is inside.

Yes YouTube seems to be becoming progressively enshittified. One specially annoying new feature of the YouTube app ( as opposed to the browser version) is that the algorithm gives you different results each time you enter a given search criterion. I use it to practice singing and often can’t get back to the recording I found last time. It insists on giving you a new selection. And the bloody ads get longer and more intrusive. Now I find if you stop at the end for more than 30 secs or so, a bloody ad will automatically start playing, even though you haven’t touched anything.

YT has a setttings menu. You can shut off watch history. In theory, this would enable a duplicate result of a previous search. IIRC you click on the little gear icon and select Manage History.

1 hour ago, TheVat said:

YT has a setttings menu. You can shut off watch history. In theory, this would enable a duplicate result of a previous search. IIRC you click on the little gear icon and select Manage History.

So the irony is that, to get back to what you watched last time, you have to deactivate the “history”?

Brilliant! 🤪

7 hours ago, exchemist said:

So the irony is that, to get back to what you watched last time, you have to deactivate the “history”?

Brilliant! 🤪

I am uncertain on this seeming paradox (the kind normally encountered in a story from Connie Willis' Oxford Time Travel Series), since I don't login to YT or use their app. It is possible that someone using the app can easily access a history list through their settings menu (which I think, in the app, you access by clicking on your face icon?). When I use it, anonymously and on the web, I sometimes encounter your problem, the same search yielding different results. The web version still gathers the history somehow, and then misguidedly "thinks" you'd like something a little different this time. Unless you deactivate history - then it will not react to your device address or anything in some sort of cache.

One way out of this morass is toa just use the clipboard. The clipboard will store that URL as long as you want provided you use the pin option. Assuming you have something comparable to Chrome's clipboard.

7 hours ago, TheVat said:

I am uncertain on this seeming paradox (the kind normally encountered in a story from Connie Willis' Oxford Time Travel Series), since I don't login to YT or use their app. It is possible that someone using the app can easily access a history list through their settings menu (which I think, in the app, you access by clicking on your face icon?). When I use it, anonymously and on the web, I sometimes encounter your problem, the same search yielding different results. The web version still gathers the history somehow, and then misguidedly "thinks" you'd like something a little different this time. Unless you deactivate history - then it will not react to your device address or anything in some sort of cache.

One way out of this morass is toa just use the clipboard. The clipboard will store that URL as long as you want provided you use the pin option. Assuming you have something comparable to Chrome's clipboard.

Yes some possible solutions there. Or else use a long enough search string to identify the performance I want unambiguously. Luckily, for the carol service most of the stuff I need is already on a playlist assembled by the choir director, so I can just reload the playlist each time. So it's just the stuff for each Sunday I have to deal with. This kind of thing, which I find needs a bit of practice as my sight singing is a bit shaky (though neumes are easier than regular score, I find):

On the business of a new ad stating up spontaneously if I leave the video for 30secs after it ends, my son wondered if was because I had autoplay enabled, so that it was moving on to the next step down the rabbit hole that the algorithm was trying to lead me down. But I've checked and I have autoplay disabled so it can't be that. Must be just further enshittification: greedily desperate for that last tiny bit of ad revenue, at the expense of the user's experience. Arseholes.

6 hours ago, exchemist said:

This kind of thing, which I find needs a bit of practice as my sight singing is a bit shaky (though neumes are easier than regular score, I find):

My wife is the music director for her church, and occasionally deals with Gregorian chants in the square-shaped neumes. Mainly for antiphons. But she converts to modern notation to them print them off. I guess neumes have ways to sorta imply rhythm, but standard notation makes it easier for her congregation. Chant, if I understand it, being unmetrical has rhythm loosely defined or implied by the words of Latin text (and basically, in her church, by watching the lady's hand gestures). Or they use Solesmes method. In any case they use a fluid and flexible lengthening of equal note values as chant calls for. I'm impressed at how a choir can achieve that fluidity so well in synch.

1 hour ago, TheVat said:

My wife is the music director for her church, and occasionally deals with Gregorian chants in the square-shaped neumes. Mainly for antiphons. But she converts to modern notation to them print them off. I guess neumes have ways to sorta imply rhythm, but standard notation makes it easier for her congregation. Chant, if I understand it, being unmetrical has rhythm loosely defined or implied by the words of Latin text (and basically, in her church, by watching the lady's hand gestures). Or they use Solesmes method. In any case they use a fluid and flexible lengthening of equal note values as chant calls for. I'm impressed at how a choir can achieve that fluidity so well in synch.

Well as you say the rhythm is not fixed as in modern music but is fluid, corresponding more or less to speech. However there are notes you lengthen, indicated by bars over them or dots after them, or by what I call a "resistor" in the note after the lengthened one. The pitch of each note is easier to read, I find, as it is just a 4 line stave and the intervals are normally simple, without accidentals (there can sometimes be B♭s). It's quite hard to sing really well, but being so ancient it has a soothing, timeless quality that seems to connect one with those who have gone before, through the centuries. I like it as a musical exercise, too.