What's the difference between AI Art and Gen Art?
Here are the differences taking into account the current NFT art trends...
Today, I would like to write about the differences between these two art styles. It all started on Twitter when I created a poll and wanted to understand the general perception about which one will be more popular in the future.
Regarding of the results, it became clear that there is a big confusion about both styles, processes, and their differences.
I came up with these definitions after seeing multiple artists create works and collectors seeking them. Those might change in the future, as technology evolves, but this is where we are right now.
What is AI Art?
First of all, let’s define artificial intelligence (AI) in simple terms.
It is the possibility to simulate human intelligence by machines.
Over the past decades, we’ve seen a massive improvement in artificial intelligence systems, due to having access to more storage, faster processors, cloud computing, and, more importantly, immerse amounts of data sets. Data sets are key since they are the gasoline for artificial intelligence systems.
AI has many applications and a vast amount of techniques that have their own pros and cons. From autonomous systems (like self-driving cars), to chat and voice assistants, visual detection, natural language processing (which is basically the capacity to understand languages and produce text outputs, answers to questions, or even generate text), and image generation. The last one, has been widely used over the past year to generate art and is called AI art.
So how does it works?
I won’t go into the specifics, as these technologies are quite complex, variants, techniques and they evolve constantly. I would just share in simple terms how they work.
Basically, in 2014, Ian Goodfellow and his colleagues create a framework called generative adversarial networks or GANs which is a subtype of neural network. In a nutshell, this technology can take data (training set) and generate a new output that contains similar characteristics. So, if you provide photos in the training set, it can generate similar photos as an output. At the beginning, you needed to technically understand how this neural networks worked, set up your own models (adding the images) and ran all the simulations by yourself. This meant that only those with the technical knowledge had access to it.
Over the last year, some applications have taken this process to another level, making it very easy for anyone to generate such outputs without the need to create their own training set. They are Midjourney, Dalle and Dalle 2. Additionally, there is a platform called GPT-3, which works in a similar fashion, but produces text outputs (it can write books, poems, jokes, stories, whatever).
To use these platforms, users just provide a prompt, or a short text. You can control the parameters and the outputs by changing your prompts. This is why sometimes this kind of art is given the name “prompt art”.
Although some “disregard” this art practice, in reality, there are many tweaks and details that make some artists stand out and take it to another level. Here are some challenges that make some AI artist
Coming up with their own training sets (their images, photographs).
Mastering the prompt game to perfectly guide the AI. As you practice and study all the possibilities, you get much better at controlling the outputs or at least, guide them into your own visions.
Come up with your own style. As the outputs can be wild, figuring out how to create your own visual style in a consistent basis might the hardest thing to accomplish by an AI artist.
What is Gen Art?
Now that AI art is clear, lets jump into gen art or generative art.