David Kitching | KitchingArt

AI is developing at a breakneck and frankly terrifying speed. The debate about whether or not it's capable of becoming conscious is raging, but certainly the possibility is not being dismissed outright. But in order to compare a 'conscious' AI with a human, at least two other criteria need to be considered, and definitions are crucial.

This is a hugely complicated debate and looking into it is like staring into an endless rabbit hole. This isn't the place for it but if you want to explore the questions it raises, may I suggest some online research? I confess I used an AI, after spending an hour trying to do it maually... I use the European AI (https://chat.mistral.ai/chat) as it's covered by the EU General Data Protection Regulation.

So let's content ourselves with some definitions. Although even these are come in various forms.

• Consciousness: The state of being aware of and able to think about oneself, one's surroundings, and one's experiences.
  This is the one about which debate currently rages. Some argue that there's tentative evidence that some form of consciousness may be emerging. Although I cannot personally see how an awareness of surroundings could come to pass as this requires an understanding of the wider context and how everything relates to one another.
• Sentience: The capacity to feel, perceive, or experience subjectively (e.g., pain, pleasure, emotions).
  Whether such capacities emerge in the future, I cannot imagine, but I'm pretty sure they're not currently present in any AI model.
• Sapience: The ability to apply wisdom, deep understanding, and discernment—often associated with advanced reasoning and judgment.
  There's so much tangled up in this one, including ethics, empathy and insight, that it's clear that current AI cannot have this quality.

Which is why current AI models require training based upon human created art such that any output is derived solely by style and content that is originally created by a human. It knows nothing of the world, and nor is it equipped to interpret what it does know with any meaningful discernment. So how could it be capable of original creativity that reflects human discernment such as emotion and aesthetic qualities? It's incapable of original creativity. It's just a machine that, very cleverly, can redeploy in a modified form, what already exists.

I'm not qualified to even guess whether or not AI will become conscious, sentient and/or sapient. I do know that the companies developing AI say that Artificial General Intelligence (AGI) is only a matter of a few years away. Whether the three qualities defined above are things that simply emerge from intelligence is a whole other debate.

I'm a photographer rather than a painter. Photography is an art form that has undergone several major technological transitions since its birth. The early days, from plate cameras through to early SLRs allowed plenty of creativity - just compare Bill Brandt with Ansell Adams for example - but even the earliest Victorian photographers used to experiment with blending two or more photographs together to create a unique image of what was impossible in reality.

Then in the 1990s, the digital camera arrived. Along with the digitisation of communications technology and the emergence of relatively cheap computers. A digital picture created on a digital camera feeds straight into a digital computer and can be instantly manipulated and blended with other images quite easily. I have little doubt that those early Victorians would have loved this. And Adams and Brandt would have relished the ease with which the tonality and contrasts in their pictures could be so easily managed rather than having to rely on the complex, slow and awkward manual mechanical and chemical methods that they used.

But now we have yet another new technology kid on the block. AI. Manipulating images on a PC, in Photoshop for example, is actually quite a complex skill and can take years to perfect. You already know how easy it is for an AI to respond to a simple text message and 'comp' together a new picture by aggregating pixel data from unknown, multiple other picture sources. This is not human creative art. We discussed this at the start. 

However - and this is where I'm going to draw a definitive line that, in my opinion, allows for the use of AI in the creation of art:

I do use AI to create some of my pictures. I use it in three ways:

1. Enlarging (or reducing image sizes. I provide a printing compnay with a file that's 9000 pixel on its longer length and this is appropriate for both large canvases and smaller pictures. Enlarging pictures is an absolutely unavoidable part of the process of putting a picture to print and AI is phenomenally good at doing it, avoiding the risk of pixellation or blurring that came with older techniques.
2. Sharpening oo softening images slightly, should some degree of this be required. This has been regular practice in digital imaging since the 1980s. 
3. Blending two or more original pictures together. Again, this practice dates back to the Victorian era. But AI, offering as it does controls over various aspects of the process, does it far better than even the very recent methods using software like Photoshop and manually manipulating the data. 
   I use it by taking two or more of my own photographs, or vector art creations, and asking AI to blend them together, applying a range of variables that I select as I do so. The result can only have derived from my own creative work.. my pictures. Nowhere else. No one else's work is involved in any way. The possibilities are endless, and whatever I set out to achieve, the process can be serendipitous with unexpected results suddenly emerging that inspire a new direction for the picture. So the process is inherently creative as well as powerful.
   Blending images with AI amounts to a new creative tool in its own right. And in fact the results often comprise a modest percentage of the overall workload because the resulting picture simply forms a third, new, picture. And that's when the long, arduous work of cleaning up and adjusting the result starts, and manually editing features, colours and tones. It doesn't always, but it can take many weeks, even months. And once I arrive at that point where I feel ok with the result... when I sense that resonance that I mentioned earlier, then I stop.

And I guarantee that the resulting picture will be completely unrepeatable. By this point, hundreds of edits and changes will have been applied and there is just no way that I could possibly remember what they all were or the order in which they occurred. Literally impossible.

So I respectfully say to those dedicated and wonderful painters out there for whom I have so much respect and who have skills and talents that I simply lack... an oil painting could be created again by a talented forgerer, but an accurate original copy of some of my work could not be made. Apart from the fact that manay prints can be made, the original piece could not be recreated from scratch, even by me.

This article started by echoing the righteous condemnation and fury of artists - photographers, painters (and writers) - who object to AI copying aspects of their original art content and style, enabling anyone to generate a picture and then calling it their artwork. Even if they don't claim credit for a picture, it is still objectionable that original work be highjacked in this way without the original artist's permission or even knowledge. 

However, to apply a blanket condemnation to the use of AI in creative processes makes no sense. AI is a tool that can be used in a number of useful and productive ways. Personally, I'm no fan of the way AI is being deployed, with caution thrown to wind as to the implications, in everything from employee recruitment to driverless cars via autonomous weapons. I think AI is extremely dangerous. But within the constraints of the topic of this particular discussion, I am finding it to be a powerful ally in expanding what I'm able to do with my own photographs and pictures. Specifically used for well defined purposes that I control and that exclusively use my own material.

So, yes, I think there is a place for AI in the creative process. The problem I suppose is that it's not possible to stop people abusing it. As is so often the case with technology, it's not the tech itself that's the problem, it's how people decide to use it.