Just a few years ago my thoughts about ai generated art were like: “well at some point we will get something valuable out of this”. Nowadays, we use AI actively for creating art pieces. With AI you can create almost any piece in any existing style. You can generate works in the styles of different art movements starting from medieval art, and to digital art.
So currently we are talking about a new revolution in the media world.
By the way, all of the images used in this article are generated by me, so they are my property. Strange to think of that as something that belongs to me since I did not photograph or paint it.
AI makes new works based on the incoming information – so do you!
You take photos based on your previous experience same as AI creating images based on input information. The most important difference is that you understand what you’re doing. AIs, at the moment, are still quite basic and don’t have any understanding of the concepts. These are narrow AIs or some people prefer to avoid the term AI and use the term “machine learning” instead. That’s because Intelligence is quite a vague term, and it means different for different people. But probably the important part of it is self-awareness and expectation.
While you can create something new using an AI generator, you can’t create a new style or genre. It’s hard to call creative, the works that are generated by AI. They all look more or less cliche because, again, AI doesn’t make associations with different concepts, but gives the output based on the input information about the pixels. When you tell an AI to generate two turtles keeping hands in outer space it will recognise the words and give the output. But it won’t understand that these are the turtles in outer space.
It might be a matter of time before AI will truly learn how to get some common sense understanding but that will take a lot longer.
According to what Melanie Mitchell said, most probably, it will take more than 100 years until we will get to AGI (artificial general Intelligence or human-level intelligence).
These predictions will, of course, never be exact, but I trust them more than my own dilettante opinion.
At the same time the debates are still going on and some say this is a matter of time since the AI that I used to generate the images in this post are only 2 times less powerful than the human brain. And the scientists might only need to find the right approach to teaching AI in the right way.
So even if AI does not have common sense, AI generated art progressed greatly within the last few years. Neural networks of AIs are just about 2 times smaller than the human brain. Yes, it’s probably unconscious. But do we need consciousness to get the AI generated art? And also what is conscious?
There are 2 theories about consciousness:
1 – Even the light switch is conscious just on a very low level. And your smartphone for example is a lot more conscious.
2 – The consciousness comes together with self-awareness and an understanding of common sense. Then cats can’t understand that they are seeing themselves in a mirror.
What art we will replace with AI generated art
Of course, you need to first define what is “art” and what’s the purpose of photography and art.
I believe that people will still value human-made art but that will be relatively high-quality art. All of the manual work in the media industry will be replaced. Simple stock images, some landscape, product and most of the commercial photography… we will replace first. AI generated art is good enough and a lot cheaper so small businesses will choose AI for effectiveness.
Some of the concept art will be replaced too, like the pictures where a single man is standing on a giant robot head in a post-apocalyptic scene.
Aside from that some of the game art and design too. I have designed and painted some game levels and I would say sometimes I would generate a shelf with books rather than painting that manually, focusing more on finishing and designing the whole level. In that sense, AI generated art could speed up the process of creating.
There are no boundaries for AI, it can develop on and on.
The human-made, more complex type of art with a stronger context will stay with us. Portrait photography for example will stay in demand as well as wedding, event and documentary photography. Sometimes people want to have actual pictures of them. It’s also bound to the memories and experiences that people have when being photographed.
AI generated art is already being used for some of the book covers or packaging designs as well as some adverts and blogs. AI will replace all the manual work. For other fields, machine learning will help to speed up the workflow like the “content aware fill” tool in photoshop.
AI generated art is why I will never come back to Shutterstock
I strongly believe that stock photography will die quite soon. I started uploading to Shutterstock 3 years ago and sometimes I thought about doing it seriously, full-time. Every time, after 2.5 weeks, I realised that I don’t like stock photography and didn’t do it continuously. I also didn’t like to depend on some services and create the most selling content for Shutterstock. I always wanted to create artworks for myself.
Recently I found out that Shutterstock is also working with AI-generated art, maybe this is not news to many people. But I have learned it recently and this makes it pointless to work in the stock photography field.
Why AI will ruin stock photography market
Let’s be honest, stock photography is quite basic and 95% of it is not very creative. It’s a lot more effective to generate an image using AI than to pay a photographer. Big micro stocks will be the first to start working with AI instead of photographers.
Why I suggest not wasting your time uploading to stocks nowadays.
You might think you will make some extra money. But to develop on a stock you need to upload at least 3000-5000 good images and a good part of these should be with models.
So stock photography requires dedication as if you’d have at least a part-time (~20h per week) job for a couple of years until you will actually earn from it. This is not a bad job when you don’t take AI in mind.
For example, I haven’t earned anything from my blog although I’m writing it continuously for over 9 months. From a stock photography market, I would get at least 100$ within this time investing the same amount of effort.
But the main problem is that in a few years such photography will be greatly replaced by AI-generated images so while you’re uploading to the stock your income will sink. Unless you become some AI stock generator and work for even cheaper.
Product photography is getting gradually replaced by AI.
Probably these are bad news for product photographers. This is a basic enough type of photography that we could replace. Some companies already use it to create merchandise examples.
But imagine this, you create some new phone case or a well-designed bottle of perfume. Then you take a few pictures of that on a phone and upload that to an AI generator. After you tell the generator to make some dramatic lighting or floating compositions with it. And in 10 seconds it’s ready and almost for free. I’m pretty sure that smaller brands will start using it to cut down the costs.
Of course, high-end product photography will not be replaced for a while because that type of photography requires a lot of creativity too.
Why AI is not a threat to photography
Photography has a certain purpose. It gives the best representation of the world that we see with our eyes. Physical world. AI can’t do it. It can very well illustrate something, but it doesn’t have a connection to the physical world. Aside from that if you look at works of real artists photographers or painters, they are incomparably better than AI generated art. They are better conceptually and are more stylised. That’s what AI lucks – the understanding of concepts.
Photography, art and design will slightly change their purpose.
As it happened when photography has just appeared, some of the draughtsmen and painters lost their jobs. No one wanted to order a portrait or architectural painting if they could get a photograph faster and cheaper. Photography did not really replace painting and drawing but changed it into a more creative discipline. It was no longer enough to simply copy the real world, they needed to express some emotions and feelings.
The same will happen with AI generated art coming into the industry. It will be simply more effective for basic needs and photography will have to become more inventive and creative.
Design and painting will also change even more. Why would you need a designer to make a simple ad or label if you can generate it? Even simpler we already have templates for all the basic stuff but AI will only strengthen it.
What AI can’t do
As I said before, AI can’t create new art pieces. If you give it a more abstract and hard request, it can’t process it correctly. For example, I told AI to generate an image with the request “the flow of time” and it gave me that:
Giving the AI a complicated request will leave you with some strange arthouse results.
Or even creepy
Sure with AI, you can get quite bizarre results but you won’t have as much flexibility as when creating art yourself. Still, Stable diffusion might give you some good AI generated art but DALL-E for now keeps giving me quite low-level art.
Think Of The Future Where AI Generated Art Is Present.
Looking at all of these technologies I would say we humans need to advance our skills in the creative and more complex fields since all the manual work will be more and more replaced. It’s obvious, yes! But with the rate at which everything is changing nowadays I would consider it now and as soon as possible.
For example, I’m continuing to do food photography simply for money. That kind of food photography is not creative at all. But if there will be some software that is able to help people create images themselves then I will lose at least a good part of the jobs. And AI generated art is able to replace me, it’s just a matter of time.
So don’t wait for something like that to happen in case you are doing some manual job in the media world. Get some courses to level up your skills and creativity and then AI will help you to create art instead of replacing you.
Here is some more AI generated art
At least for now, the images in this post are lower quality than mine. You can check out my other posts to see my works there:
If we talk about the style – AI generated art is something rather pop than stylised but that’s obvious since it takes the input from all the images that were provided. When I specify the style more – it can create something in some artistic style as well as the style of a specific artist. I wonder if people will create some artistic AI that will generate some stylised images in its own style.
Thanks for reading till the end, I hope this article gave you some insight on what to expect from AI generated art and saved you some time that you would dedicate to a field that doesn’t have its future. If you liked this article, please subscribe to support me in what I’m doing.
