AIRTIST—Matrix’s Great Venture into AI Art

What is AI art?

In recent years, as a novel form of art, AI art has caught more and more mainstream attention. AI is short for artificial intelligence. If computer-generated imagery can be called digital art version 1.0, and art created by people on the Internet is version 2.0, then the emerging AI art may bring us into the third iteration of digital art. From 1.0 to 3.0, digital art is constantly upgrading and revolutionizing itself. As we have witnessed, CGI-based digital films, anime and online games are some of the most powerful driving forces of digital art. Internet art based on various interaction mechanisms are also gaining huge traction.

AI is the technology that makes machines capable of performing cognitive tasks as humans do. These include playing chess, image recognition, translating, searching for the best candidate among heaps of resumes, etc. We can extend this concept to art by having computers learn the techniques and styles of classical artists and create works that pass as from the same period in the eyes of art history scholars.

In fact, this is similar to the idea behind the famous Turing test. If artworks created by AI after training can pass as works from a certain historical period in the eyes of art history scholars, and these works are not simple modifications of their source data, then these AI artists have passed the “Turing AI Art Test”. All this sounds simple enough, so we can use it as our fundamental definition of AI art. Under this definition, AI art is any artworks considered legitimate classical or contemporary art by experts in the relevant academic fields.

In today’s world technologies are developing at a breakneck speed, and the definition and boundary of AI art are also changing and becoming clear. On the one hand, people are beginning to try using AI as a tool to assist humans in art creation. On the other hand, people are also using the latest technologies, such as generative adversarial networks to help AI understand art in order to enhance their mastery.

Portrait Of Edmond de Belamy: A Milestone of AI Art

In October 2019, Portrait Of Edmond de Belamy (2018) the first-ever AI artwork was sold for 350,000 USD after a heated six-minute auctioning, smashing previous estimations. This transaction was settled at 432,500 USD including commission, which was more than 43 times the estimated price.

Starting from this, more and more AI artworks have entered the global spotlight. In March 2019, Sotheby’s sold Mario Klingemann’s Memories of Passersby I for 51,012 USD. In the spring of 2018, HG Contemporary Art Center held the first gallery exhibition dedicated to AI artists in Chelsea, Manhattan. Among the exhibits was a work called Faceless Portraits Transcending Time, a collaborative work between AI and Ahmed Elgammal, professor of computer science at Rutgers.

Renowned art institutions and collectors alike are closely following this new form of art. Marisa Kayyem, director of Continuing Education at Christie’s Education, said: If we look at what is happening in the art world from a broader perspective, instead of just fixing our attention on sales, we will realize the things that are happening are powered by strong momentum and backed by large institutions. More and more collectors are getting used to AI art.

Many professionals in art-related industries still hold a degree of suspicion towards AI art, thinking that the word could be misleading or has too narrow a reference. Like other artists who are also programmers, Klingeman prefers to use the word generative art, which refers to all works created using algorithms. Klingeman’s works have been put on the auction block at Sotheby’s, while the origin of generative art can be traced to the late 1950s.

Jason Bailey, founder of Artnome, said: AI art is actually a word made up by media in the recent three to five years. He thinks that using this word will mislead people into thinking these are art created by machines. The majority of artists I know don’t like being called AI artists, he said, but like it or not, the word is quite commonly used.

Although Edmond de Belamy is the most well-known AI artwork, for people getting to know this art form for the first time, the work might be a bit distracting. Edmond de Belamy was created using generative adversarial networks (GANs), which was trained using a set of artworks to learn to create a similar style and use all the knowledge to create something new while retaining certain elements from the original images.

NFT and the Metaverse: New Opportunities for AI Art

2021 was the year NFT gained serious traction along with everything related to the Metaverse. Top-notch teams from all over the world have entered the scene, and the Metaverse also saw the very first influx of new users. Although the hype has died down a bit since then, this doesn’t change the fact NFT is getting increasingly accepted by the general public, especially with the setting up of infrastructure such as Opensea, Huahe, Jingtan, etc. The enthusiasm with which people can accept something revolutionary like NFT is beyond expectation, while profile picture NFTs are leading the industry in terms of creativity and growth speed.

Some eye-catching success stories of NFT include Mike Winkelmann’s Everydays: The First 5000 Days, which sold for an astonishing 69.3 million USD, setting a record for digital artworks. Jack Dorsey sold his very first tweet for more than 2.9 million USD. Even more insanely, British contemporary artist Banksy’s Morons was bought at 95,000 USD, and the original work was destroyed intentionally after being converted into an NFT version, which sold for 380,000 USD.

Most people would find it difficult to make sense of all this. But the fact that NFT can attract so much hot money is all thanks to the hype about the Metaverse. In March last year, Roblox was listed on New York Stock Exchange and immediately favoured by market capital. This incident popularized the concept of the Metaverse like never before, and as the fundamental technology of the Metaverse, NFT has come to be known as the key to the digital world. If the Metaverse is a parallel universe to ours, then NFT is the bridge between the real and the virtual worlds.

AI art is a fast-growing new industry with limitless potential. The popularity of the Metaverse and NFT will create more opportunities for everyone involved.

What is AIRTIST?

AIRTIST is the combination of two words: AI and artist. As the name suggests, AIRTIST is an AI artist, a distributed program that uses AI algorithms for art creation. We use deep neural networks to create AIRTIST, and every AIRTIST is operated by an independent deep neural network. To guarantee AI artists and their works are decentralized, these deep neural networks will be deployed on the blockchain.

After an AI artist has been created, users can use distributed computing power to run the DNN corresponding to the AI artist to create artworks.

Depending on the type of art, AIRTIST will create four kinds of AI artists: PAINTER, MUSICAIN, DESAIGNER and WRAITER. AI artists will branch out into more art forms in the future.

Why do people need AIRTIST?

AIRTIST will transform the current art NFT market in three ways:

1. Solving the problem of trust between online and offline parties which troubles conventional art NFTs:b Conventional art NFT are created by human artists in the real world. Therefore, both the digital artworks and their artists are centralized. The transactions of these artworks invariably take place at centralized exchanges, which is at odds with the ideal of decentralization that blockchain represents. On the contrary, AI artists’ works, from their creation to their storage and transaction are all decentralized, which perfectly solves the problem of trust during NFT trading.

2. Lowers the entry barrier for the NFT market: In the current art NFT market, artworks are created by a minority of people, while the average person has no way to own an artwork except by paying the costly price to buy one. The arrival of AIRTIST allows everyone to own his/her independent artist. Besides buying and selling, oridinary users can create their own art NFTs, which will find their way into the market. This is expected to make the current market size many times larger.

3. Bringing deceased masters back to life: Through machine learning, AIRTIST can master the styles of deceased artists such as Van Gogh or Monet. The world will marvel at new masterpieces from these deceased giants.

Work by an AI Artist trained using Vincent van Gogh’s works

AIRTIST and Matrix

AIRTIST has three core components: AI artist, art creation and art circulation (minting NFTs, buying and selling). Among these, AI artists and art creation will happen on the Matrix Mainnet, while the circulation part will take place on Ethereum.

Why choose Matrix?

Matrix AI Network is a highly efficient blockchain empowered by AI. Here are some of Matrix’s features that will benefit AIRTIST as well as AI art creation.

1. Compatibility with IPFS: Distributed storage is the fundamental storage facility for NFT, and the Matrix Mainnet is already perfectly compatible with IPFS. Through future protocol updates, Matrix will also support Filecoin and other distributed storage platforms.

2. Distributed Computing Power Platform: Matrix provides a distributed computing power platform, which will guarantee sufficient high-quality computing power for the creative work of AI artists.

3. Algorithm Authentication Platform: The Matrix Mainnet supports algorithm authentication and minting algorithms into NFTs, which protects the intellectual property of AI artists who train DNN.

How to create an AI artist?

To create an AI artist is essentially to train a DNN, which uses the same methods as training generative adversarial networks. Here we need two deep neural networks. One is called AIRTIST-DNN and imitates the creative process of artists. The other is called Discriminator and serves to improve the quality of the art. The relationship between the two networks is illustrated below. On the bottom left corner is AIRTIST-DNN, and Discriminator is in the middle. An AI artist needs to be trained before it can create art. The two networks are deployed on the Matrix Mainnet using distributed computing, and we need a group of artworks of a certain style. Before training begins, the weighting of the two networks are set to random. AIRTIST-DNN is built using a generative model, which can mirror a set of random numbers onto an artwork. Of course, in the beginning, the results may not look like art by any means. But in the meantime, Discriminator will also start training, using preselected art samples and works created by AIRTIST-DNN as sample data. The target of training is to be able to tell which one is the real artwork and which one is generated by AI. In the process, the weighting of the two networks is constantly readjusted to get the statistical distribution of AIRTIST-DNN-generated data as close as possible to that of real art. This training process can be seen as a game where two players compete with each other. The quality of generated works will match up to real ones when Nash equilibrium is reached.

How to create AI artworks using AI artists?

Let’s use the process of AI artists creating a painting as an example to illustrate how AIRTIST works.

A fully trained AIRTIST deep neural network will have multiple layers of neurons, each with large quantities of weighting parameters, which reflects the statistical distribution of specific artworks. To create an artwork, first, we need to enter a theme. After semantic analysis, the theme will be converted into a group of preliminary parameters to be sent to AIRTIST-DNN. After the data input, neurons of the hidden layer will calculate the data and weighting parameters level by level. Finally, the output layer will receive the final result, which is the end product of the AI artwork. In fact, a deep neural network is an all-powerful tool that can mirror data input onto any data output. The process of training is to adjust the mirroring function of the network to match the target art style, and the creative process is to generate a set of data (i.e. the artwork) based on the distribution style reflected by network topology and weighting parameters. As a deep neural network will have millions of weighting parameters, there will be a degree of randomness in the reflection of specific distributions. This randomness is a manifestation of the artist’s free will.

After a work has been created, it is authenticated through blockchain and saved on the distributed storage to be a digital asset.

Why is AIRTIST important for Matrix?

Both AI art and NFT are practical use scenarios for Matrix. Before the AIRTIST team joined, Matrix had already been planning to build its own AI art platform, which attests to how important AIRTIST is for the Matrix ecosystem.

1. The First Important Use Scenario for MANIA

Every AI artist is an independent AI algorithm model, and AI artists themselves can be minted into NFTs to go into the market. All these are thanks to MANIA’s innovations. As the first major application of MANIA, AIRTIST’s impact on the AI art scene and the NFT market will likely create more use scenarios for MANIA. The integration of AI models and NFT will spark more innovations in the whole industry.

2. More Demand for MANTA

Whether it’s training AI artists, or creating art using these AI artists, distributed computing power will be necessary. AIRTIST will be one of the earliest projects to be granted access to computing power from MANTA. AIRTIST will create more demands for the distributed computing power produced by Matrix, and this in turn will make it more profitable for Matrix miners to output computing power.

3. More Liquidity for MAN

Matrix services and products offered through MANIA and MANTA will be paid for in MAN, so the popularity of AIRTIST will create more liquidity for MAN.

4. A More Mature Matrix Ecosystem

The integration of AI and NFT plus the introduction of AIRTIST under DAO will diversify the Matrix ecosystem and bring about more opportunities. Creations of artists on AIRTIST could be packaged into an AI service offered on MANAS. There could even be professional AI art critics trained by MANTA.

The Future of AIRTIST

NFT is being accepted by users at an unprecedented speed. OpenSea’s daily active users hit a new high at the beginning of this year. According to Dappradar, the volume of transactions on OpenSea reached 4.491 billion USD in January 2022, up 102% month on month. The number of transactions rose to 2.8163 million, a 95% increase month on month. Its daily active users registered 46,800, achieving a historic high and a 86% increase month on month. This data shows that more and more users are starting to not only accept NFTs as digital assets but also trade and hold them.NFT is being accepted by users at an unprecedented speed. OpenSea’s daily active users hit a new high at the beginning of this year. According to Dappradar, the volume of transactions on OpenSea reached 4.491 billion USD in January 2022, up 102% month on month. The number of transactions rose to 2.8163 million, a 95% increase month on month. Its daily active users registered 46,800, achieving a historic high and a 86% increase month on month. This data shows that more and more users are starting to not only accept NFTs as digital assets but also trade and hold them.

After over a year’s fast growth, the NFT application ecosystem has developed into six main areas: collectibles, artworks, game equipment, virtual lands, practical NFT and financial NFT. Of them, collectibles, artworks, game equipment and virtual lands account for the largest market share.

Based on data about all NFT-related niches, collectible NFT has become the predominant use scenario. The transaction volume of collectible NFT this March has reached 995 million USD, and the volume for the first quarter registered 4.134 billion USD, up 67% from the 2.461 billion USD of the previous quarter. The hot collectible NFT market is the primary force in helping the NFT industry recover since the beginning of the year. Starting from July last year, the turnover of collectible NFTs has been sharply on the rise, and its proportion in the entire NFT market rose from 28% to 84% by March.

AIRTIST will effectively lower the entry barrier to NFT and greatly expand its market size. Thanks to AIRTIST, the NFT market will welcome a new round of explosive growth, and AIRTIST will benefit from being the pioneer of a new age.

Users will have diverse demands for art in the Metaverse, and the majority of these artworks will be created by AI. This will mean huge demands and growth potential for AIRTIST.

The arrival of AIRTIST could mean a crisis for the current art market as well as some artists, but in the eyes of many event organizers, institution managers and academics of the art industry, AI will create countless opportunities by being a new player in the game. Human artists may engage in a battle of imitation and counter-imitation. Some human artists will begin to use AIRTIST as an innovative art creation tool to achieve unique styles, just like skilled photographers are able to get creative visual effects in their photos by adjusting the technical parameters of their equipment. Who is to say this won’t be the beginning of a dozen new schools of art?