I can't find the article, but several years ago I read one about a basic point-and-click digital camera given to a couple amateurs and a couple professional photographers. The difference in outcome was astounding, with the professionals taking shots that would look at home in National Geographic and the amateurs taking shots that would look at home in a home album, perhaps.
Given the same base tool and opportunity, the obvious difference comes down to the application of experience and skill. However, now that cameras of all kinds have gotten better, that gap is closing as the digital algorithms applied by the software mimic many of the skills of the professional, so the amateur can get a similar result. However, it also depends on conditions, so the camera can't mimic things like being in the right place at the right time, nor getting the right angle of a shot, so there is still a difference.
However, while the quality of images has gone up, the value of photography has gone down, because the cost to take a good image has been radically reduced. It is not just the cost of equipment that has come down, but because of the coding, the cost of getting the skill required has also come down, because there is no need to actually experience many of the conditions, the camera simulates it for the photographer instead. Essentially, an image is taken by an out-of-body, photography cyborg, a partnership between human and machine.
While that is another topic to dive into later, this post is about the acquisition of skill and the devaluation of skills as they become easier and cheaper to acquire. Essentially, the easier a skill is to get, the less marketable value it has for those who have it, and the higher the level required becomes in order to give it value. Take kicking a football as an example; almost anyone can do it, but relatively few can earn from doing it.
In terms of photography for example, there are now millions of people on Instagram who claim themselves to be photographers or have an interest in it, even though they do not possess many of the skills required themselves. What this means is that while they might feel that they have a skill, they actually do not. You would not want these people to shoot your wedding photographs.
Applying this to the discussion to content created by AI, going back to the first paragraph and the camera example, what happens to content when all content is largely the same? The reason it is the same is because it is all largely coming from the same centralized algorithms and the cost to do so is very low, so there are going to be a lot of people putting in their terms and getting an output that is similar to a lot of other people. In fact, there is no need for the people at all, because an algorithm can create its own generative terms and have closed-loop, humanless content creation system.
However, what it is actually doing is feeding off what is already created and even though there are essentially an infinite amount of configurations, eventually without new material, it will run dry. Yet, for the human experience, this doesn't really matter, because we can't consume that much. It is similar to how humans can only see about 1 million different colors, yet develop screen qualities talking about 16 million variations. It is like making a song for humans, in sound frequencies only dogs can hear.
So, what gives content value, when everyone can cheaply create the same quality content?
This is a good question. But since we talked a little about music above, let's start there. Computer generated and assisted music has been around for decades, and tools like autotune (that twangy voice) have been fixing the skill of singers since 1996. However, while we might listen to a lot of computer generated music on the radio, would you pay to go to a concert of AI-generated music?
What about an exhibition of AI-generated artwork? Once? Twice? How many exhibitions would you go to, knowing that the images took literally seconds to generate and that all it was doing was doing was taking images from the internet, applying an algorithm to them to combine some kind of mathematical outputs into something visible?
Hmmm.
Perhaps there is value in humanness after all.
The value of skill isn't the output of the skill in most cases, it is the journey required to get the skill. And that journey is unique to the person experiencing it. Unless an AI has access to all conceivable data in my past, it cannot produce what I am producing now, in the same way I am producing it. This means that even if it does give the same output to the letter, the consumer given both versions would most likely value the one that has taken humanness applied to get the result. If they know that is.
Which takes us back to the music.
Why go to a concert if the studio recoding is better quality? Atmosphere? Sure. But there are also a couple of other things that quickly come to mind, with one of them being proof of human related. We want to experience the artist themselves and in so doing, show our support, put our time, money and effort behind them. Show them we appreciate their skill and by proxy, be part of their journey. It is a shared experience, not only with the rest of the crowd, but with the artist themselves. Would that shared experience feel the same listening to an AI perform a concert?
How much would you pay for tickets?
I paid €250 a ticket for my wife and I to go to see Adele live in concert in 2016. We made the trip to Sweden and stayed there a few nights in order to do it. Would we have paid the same to go and listen to a CD of her, while she sat there on stage? Would we pay that to go and listen to the CD of her? Would we pay that to go and listen to someone else sing her music? Would we pay that to go and listen to an AI sing music similarly to her?
The question isn't necessarily whether AI-generated art or music is art or music or not, it is more about what is the value of the output. And in general, what we value is going to be the journey, which only the human can take, even if the AI is drawing on the journeys of many humans, which they are. Similarly, it is with text, where we would feel cheated if we discovered what we thought was human-generated content, turned out to be AI-generated. The reason is that we actually value the human experience more than the mimicked experience of the AI, in the same way that an original painting is worth more than a print of that painting.
Whether it be art, music, video, photographs or text, if we have supported the content as if it was human created, we would feel cheated and swindled if it wasn't, because for the most part, we wouldn't ever support it at that level otherwise, because we value it less.
As said, the lower the bar to get a skill, the lower we value the output of that skill and like it or not, the artistic outputs of an AI is far, far lower in skill level than the equivalent outputs from a human. When an artist spends years getting the technical skills and building the experience to deliver a phenomenal painting, it is part of the shared experience that we have with the work, no matter if an AI can get the same output.
We don't want to go to a dance party and listen to a Spotify track list, we want a DJ to apply their skills. And this is where the human comes into it, because even if it is computer-assisted, we want to experience the application of the artist's unique journey and appreciate what has got them there, onto that stage, and into our minds and hearts. An artist isn't one image, or one song, they are just the outputs of their travel, their trip, like our own, that has propelled them forward into life to deliver something we value.
It is not about whether we can tell whether the output is human created or not, it is about us being able to assume that it is. Because if we assume otherwise, we will value it less, because it means that it wasn't difficult to accomplish, and we like the hero story, the ability to overcome adversity, because we can connect with it personally.
Acquiring a skill is a challenge and depending on the skill or level, not all can achieve it. And because of this, the law of economics comes into play and as we should know, in a world of abundance, scarcity still matters, it has value. While we wish that there was an abundance of food and opportunity for all, when it comes to the value of art, it is rarity that sets it apart, and AI generated anything means that it no longer is rare, it becomes a production line of mass produced outputs - cheap and disposable.
Perhaps at some point we will no longer care about whether something is scarce or not, but in a world where people believe that their unique identity is valuable because it is exclusive of others, but also groups them with others who share a particular journey, it might still be awhile away. However, we are lazy animals and love convenience, so we will likely speed the process to our own destruction, by taking the easy path and relying on the tools we create to get an output of professionals, even though we ourselves are unwilling to take the journey to actually become a professional ourselves.
I do think though, that unless entirely enslaved, human creativity that provides scarcity of experience will still get the bulk of the attention and, the value support. An AI can do a lot of things and fool a lot of people, but proof of human will become a thing and the unskilled, will not be able to compete with the skilled. Their only savior is that for now at least, there isn't the transparency of activity required, but in time, that web of trust will be able to provide clarity on who can really create, and who can only imitate the output.
When you look in the mirror of your content, do you see yourself?
If you do not, at some point, others will not care to see you at all.
You played yourself into a prison of irrelevance.
It is the journey that matters, not the destination.
Taraz
[ Gen1: Hive ]