Recently I have been bombarded with ads on Instagram and such about apps creating AI-based art. The concept in and of itself is super interesting and obviously somewhat controversial, so naturally I had to investigate further.

I didn't want to give certain dodgy apps the clicks so I simply went to the play store and found 'StarryAI'. The most famous one going viral often at the moment is Dall-E by OpenAI, but it put me on a waiting list to join so I figured I'd come back to that later. I've since activated that account but I've yet to fully explore it.

I get the initial impression that Dall-E is a lot more... tangible? While StarryAI is a lot more abstract. It has a lot of cool options, including two different AI setups, and even a list of styles from various real and historical artists to base the art being rendered on. To me, the outcome from this app seems much cooler, but I guess we'll see next time.

Here are the Prompts I gave to StarryAI, selecting random artists as I went:

1. 1,000 bears

2. Real artists' revenue being stolen by AI while musicians watch on in support

3. War of the colourful ants

4. Balrog Dog

5. Dreamy dystopian lack of freedom (here I uploaded an initial image of a street sign indicating 'no dogs'

_{My personal favourite... the weird statue of liberty and American flag really put the cherry on top here}

6. Galactic Nightmare

The reason I find these so special and profound compared to Dall-E is precisely because it is a more imperfect rendition of the things I describe. Here's a few of my favourite aspects:

The not-quite ants

And the War of the Colourful ants, as well as 1,000 bears, real artists' and balrog dog, the abstraction and distortion of the main figures is so strange, incorrect and yet so familiar, it boggles the mind.

The humans in Real Artists are... just not quite humans, and yet we can still distinguish the fact they are meant to be so, even though we can't necessarily even tell if they are one or multiple humans.

Likewise, you can clearly see the ants are meant to be ants. You can identify the wings on the flying ant, the thorax and abdomen, mandibles, legs and head are all somewhat distinguishable, and the ensuing village fire from the terrible war, even though it's not... quite... a village, and the fire is just not quite a fire at any given point, even though we obviously know it's a fire in principle.

The bears are a much clearer example of this. They are clearly not bears, but somehow, if given to an unknowing human, the chances are quite high they would identify them as 'some kinda weird bears'

It's this kind of abstraction and warping of reality that I think would be extremely difficult if not impossible for a regular human to do, as we are all imprisoned in our own realities. All our personal abstractions are all based on warping very real phenomena and experience in our lives, whereas AI doesn't have such a mental roadblock.

It's not art!!

My friend who came up with the 'real artists' prompt was very quick to assert his opinion that this is just another way AI is destroying human careers, taking away our jobs, so to speak, and reducing 'real' art to computer technology with no heart or soul to speak of.

An ex student of mine had a similar rant about it on my social media, insisting that it is the years of toiling and struggling over a paintbrush, a lifetime of servitude to the art itself which makes it a real art form. Now I can just click a button with no cost and come up with this disgusting replacement with no meaning. Art will, as he says, gradually die.

Personally, I heavily disagree with both points.

AI is stealing our jobs

Well, in many ways this is obviously true. Robots of many kinds have been doing that for generations, from car manufacturing to bomb disposal.

But it is not the fault of the AI so much as it is humanity's incapacity to continue to innovate for themselves. For example, when a coal mine shuts down, one of two things can happen: A coal miner can lose his job and become a jobless, homeless bum, or he can be re-trained in the field that was set to replace his old job such as solar panel installation, wind farm engineering, maintenance and so forth. These jobs are currently in plentiful supply, but society has done a poor job of providing people those opportunities on the whole (many industries do in fact do a fair job at this).

Well what happens when those jobs die off? This is why we require human innovation. Our generation for the most part knows little about the very concept of being a coal miner. Most in the West cannot even fathom such a dangerous, horrible and low-paying career. Well what are these whiners doing now instead? more often than not, it turns out, the programming required to make the AI that replaces them in other fields!

The new frontiers created by the previous technological advancements have no end in the foreseeable future, as long as we continue aiming to become a multi-planet species, however there is still a LOT to work on here on Earth for perhaps several generations at least.

Regarding the AI making Art, well, I think you can see there is still a LONG way to go. In music, when it comes to making music for a movie, directors have two choices. They can peruse a vast, vast library of pre-made compositions that happen to fit their project with a bit of fine-tuning. This can cost them basically $1, and most low budget projects do this exclusively, from my experience. Commercials, video Intros and so forth.

Well, Humans are the ones who innovated the concept of these royalty-free sound libraries and humans are the ones profiting from it.

But higher budget projects might want compositions that are specifically catered to their movies or shows. This is where AI is a long, long way from replacing. You would have a hard time making AI create a soundtrack which exactly depicts the emotions and timing which the director desires at any given moment. It fails to be able to take in feedback from the director to brush up, scrap, swap out and re-orchestrate every moment of music to match what's happening on screen.

If a director doesn't much care about this element of their movie, they are a bad director and should just stick to the cheap ready-made libraries. For now, AI is unknown decades away from this kind of accomplishment.

I believe the same applies to art. I'm unaware of any artists who thinks they can simply describe what they want the outcome to look like in words which an AI could perfectly render. I mean, when I asked for 1,000 bears, I wanted 1,000 bears, not 30-ish brown blobs that kinda look like weird bear-rock things'.

If an artist wants such a vague piece of art because it looks cool, then perhaps it is they who are the problem.

Art, for the most part, is not simply something you put up in a gallery, either. Most art these days has a corporate function, and corporate function of art typically takes billions of dollars worth of Research and development, analyzing and studying human emotional response, colour connotation, blending & matching according to the current status quo of a given culture (red has different meanings in China compared to USA, for example), how to catch an eye and how to make it sell.

In this sense, AI is totally useless and there is no progress to even suggest it will even attempt this kind of endeavour for decades or more.

When it inevitably does, it won't even matter anyway because we rarely simply find passion in a random piece of art. We tend to follow the art for the artists we respect, admire, empathize with or idolize. The drugs and the degeneracy, the depression or demographic that the artists' soul is inherently built from, is reason to follow, in the same way we buy a music album from a musician before we have even heard it, because their last one got you hooked.

You loved that one Dream Theater album so you looked them up, attended their concerts and bought their merchandise and eagerly awaited their new album now that their drummer has been replaced, so you can furiously debate which is better, old vs new in online forums.

AI, no matter how human-like its art becomes, lacks the very canvas of life that we are drawn towards to begin with: the human.

No effort therefore no art

I just think this is a naïve statement. Nobody said art requires effort to become defined as art. As I'm a musician I'll use musical analogies once more. John Cage infamously created 4'33", a piece of music in which the performers play no notes, make no sound, and simply sit there for 4 minutes and 33 seconds. However, the music can be performed at any length, with many versions performed 7, 9, 12 minutes or beyond.

There are death metal band covers of this 'music' and decades of controversy, making the piece extraordinarily famous in the music world to begin with. The piece screams the question 'what is music'?

You'll tend to find the more one tries to define music, the more difficult it becomes to deny that 4'33 is indeed also music. 'Organised sound', perhaps? Well, 4'33 is organised, written down. It even has 3 movements.

A car engine is organised sound.

The dictionary definition of music is 'Vocal or instrumental sounds (or both) combined in such a way as to produce beauty of form, harmony, and expression of emotion.'

Well, I can think of a lot of music that lacks form, harmony or expression of emotion so that's trash. Just take a look at many serialist pieces for a lack of form. Gregorian chants for a lack of harmony, and commercial corporate music for a lack of emotion. Not only this, but the dictionary requires 'beauty', which is strongly disagreeable at its core.

Nowhere does music - or art - require effort, time and experience. It is the outcome that matters. One artist might throw some paint on a canvas randomly for 20 seconds and sell it as a political message for millions. Others might do the same as a way to get around tax burdens. Or, one might spend her whole life struggling with a single concept before concluding the best way to express it is to throw random paint on a canvas and sell it for $20 at a charity store.

Who is to say one is more legit than the other?

So, when an AI comes along, who was created by humans, and uses human prompts to create art which is based off of a human-made database of human-made images, who is to delegitimize the outcome simply because it only took 15 seconds to render?

Seems legit to me. If I see it as art, it is therefore, demonstrably, art.

I'm sure many disagree with every word i've said, though.