NFT art created by AI algorithms brings new meaning to nonfungible tokens, but can artificial intelligence be trusted to produce a new genre of art?
Sales of nonfungible tokens, or NFTs, reached $25 billion in 2021, demonstrating that the sector is one of the most sought-after markets in crypto. Art NFTs, in particular, made a big impact last year with Christie’s reporting over $93 million in nonfungible token sales during its fourth annual Art+Tech Summit that took place this past August.
While notable, much of the crypto art scene appears to be dominated by cartoons and memes, as projects like CryptoPunks and Bored Ape Yacht Club have taken center stage. Although these projects are some of the most successful to date, a new subset of NFTs is emerging based on advanced technologies and the human imagination.
AI-generative NFTs become a new art genre
Known as “AI-generative NFTs,” these nonfungible tokens are becoming increasingly popular within the art community, along with those interested in emerging technologies like artificial intelligence, blockchain and the Metaverse. In order to create AI-generative NFTs, one would typically use generative adversarial networks, or GANs. These are algorithms that leverage computers to use data to train models to produce machine-made images resembling art.
Claire Silver, an AI-collaborative artist, told Cointelegraph that AI-generative NFT art is a relatively new genre, noting that the basic principle is that art is created in tandem with some semblance of artificial intelligence, like GAN:
“There are code-heavy options and completely code-free tools that anyone can work with. I use the latter in my work. Being able to work with an AI to bring your ideas to life is an experience like no other, it augments creativity in a way that feels like freedom, a type of play you haven’t experienced since you were a child.”
In order to create AI-generative NFTs, Silver explained that she leverages a text-to-art generator called “Eponym.” Developed by the AI-generated art company, Art AI, the Eponym tool allows users to create art based on their text of choice and then mint these creations directly to the largest NFT marketplace, OpenSea.
“Cassandra Ex Machina” Source: Claire Silver
Eyal Fisher, co-founder of Art AI, told Cointelegraph that Eponym allows for any phrase to be transformed into a unique NFT art piece that will forever be etched onto the Ethereum blockchain as a visual representation.
Fisher explained that Eponym was built on algorithms for personalized generated art that lets people create art by interacting with a computer. “Eponym is a collaborative NFT project. Users access it by coming to the website and typing any phrase or word into a text box. The AI then generates artwork based on the text entered.” Fisher added that each text prompt can only be generated once. “There is only one Eponym called ‘Bitcoin,’” he said.
“$btc” image produced by Eponym. Source: Eponym
Although AI-generative art is a fairly new concept, Fisher shared that the first Eponym project sold out overnight on OpenSea, making it one of the largest collaborative art projects created by 3,500 different artists. “This is an experiment in decentralizing art. People who own Eponyms are creators of that art and want to curate it,” he said.
While Eponym lets users create their own art NFTs, Metascapes is another project that was developed by three photographers looking to combine human expression with computer algorithms. Ryan Newburn, one of the photographers behind Metascapes, told Cointelegraph that the project consists of 3,333 rare AI-generated NFTs based on photographs taken across the world. Like Eponym, Metascapes leverges AI algorithms to create nature-inspired NFTs. According to Newburn, the first Metascapes collection is planned to be minted by the end of this month or early February.
“Ice Journey” Source: Metascapes
The AI-team behind Metascapes — which goes by the name Versus Labs — explained that the artwork in each collection is created by training data to recognize real-world images:
“We have images and labels for our photographs, which are called ‘training data.’ When it’s time to generate the output model, we put in a label that tells the model what type of images to put out. For example, ice caves and volcanoes were two categories the photographers have done work for in the past, but this wasn’t a majority of the input data, so we wanted to make sure the output contained examples of ice caves and volcanoes.”
Versus Labs added that Metascapes’ AI learns as it goes, noting that the generator that outputs data gets better over time since two models of learning are being used in tandem. “The generator outputs random noise at the beginning and the discriminator model tells the generator how to improve the output, so it looks more like the training data. This cycle continues, ensuring both models improve over time.”
Iurie Belegurschi, another photographer behind Metascapes, told Cointelegraph that as photographers, the Metascapes team chose to work with AI to generate images because everything related to NFTs and Web3 is about machines, computers and code:
“We decided to collaborate with a machine to create an entirely new world in the Metaverse. Everyone now is buying PFP avatars, but in our case, people will get a unique piece of land in the Metaverse.”
AI’s impact on artists and NFT collectors
Although AI-generated NFTs are still an emerging concept, this model has started to impact both artists and NFT collectors. For instance, the digital artist known as “Kami was Here” told Cointelegraph that working with AI has fundamentally changed the role of the human artist:
“The human needs the algorithm and the algorithm needs the human. For me, this new role meant data collection, writing code, curation, the inspiration to create a theme and, most of all, ‘coaching’ an algorithm. The process is dynamic and the outcome collaborative.” “The Cartographer” Source: Kami was Here
In terms of how generative art NFTs differ from other nonfungible tokens, “Kami was Here” explained that each result is fully unique since it is birthed from an algorithm. “Generative art explores the future of a society hardwired with human-computer interaction,” the artist mentioned.
Moreover, Fisher pointed out that accessibility and a newfound desire to own NFTs has been an outcome of AI-generated images. “Many of our users are creating NFTs for the first time, while NFT collectors and buyers are making their own creations. This is unusual, as most people in our community are not professional artists.”
Dr. Alex Alter, principle AI-scientist for Altered State Machine — a protocol that uses NFT intelligence to create smart AI agents — further told Cointelegraph that not only do AI-generated NFTs look unique, but they also bring a deeper abstract feeling to individuals. “These NFTs are truly unique in the sense that there is no single area in any of the AI works that have similar pixels. Also, in the future, people will be able to create AI artwork through DAOs and chatbot technology. This is far more than what other NFTs can do today.”
“Singularity by AIIV” Source: Dr. Alex Alter
Can AI be trusted to create meaningful NFTs?
While the potential for AI-generative NFTs is apparent, the question of whether or not artificial intelligence can be trusted to generate quality images based on text or photographs remains a concern.
As such, Newburn mentioned that Metascapes carefully curates each of their collections. “Our AI team has generated tens of thousands of images. Not all of these will be showcased in our mints. If we aren’t satisfied with the category, we strategize and retest what categories will work with each other. Our AI has learned from multiple tests.”
It’s also important to note that there are different ways to generate AI-based NFTs. For instance, Fisher mentioned that Eponym has two versions of its generator available to the public, one on the company’s Discord channel operating as a chatbot and the other as a private link that contains more complex algorithms capable of creating more advanced images. “Kami was Here” further pointed out that some AI-art pieces can take only a few minutes to generate, while others can take longer:
“There are free apps now like Wombo that can easily generate images. It’s simple for people to create. On the other hand, AI art can also take months to build and train your algorithms, collect input data and pay for processing power. AI art can also be very resource intensive and personalized.”
Technology aside, AI-generative NFTs are bound to be a disruptive trend moving forward. According to Dr. Alter, AI-generated art has already seen huge volumes on OpenSea, noting that the market will continue to grow this year. He mentioned that this will be the case partially due to the functionality of AI-generative NFTs. “In the future, people will be able to own their own ‘AI artist’ NFT (AI which can produce art) or use a DAO AI to create art together with that AI artist.”
Additionally, the rise of the Metaverse should prompt the growth of AI-based NFTs. For example, Fisher remarked that Eponym’s next project will feature interactive virtual identities where users can take their own portraits to create 3D avatars and animate them using artificial intelligence. “Our idea is to use AI that will allow for avatars to take different shapes that are compatible in metaverse environments like Sandbox. In February, the company will be introducing additional algorithms that will allow users to generate personalized avatars.”