1. Validation

Selection and validation process for artists and base image creators. Through the collaboration and fair process involving the DAO, validators, and general evaluators, the registration and validation of artists and base image creators are carried out. In the first step, when artists and base image creators apply for registration, the DAO's vote determines whether they will undergo artist/creator verification. In this process, DAO participants review the application contents and evaluate whether the applicant has sufficient qualifications. The evaluation by the DAO takes into account various factors based on each member's judgment. In the second step, the selected artists and creators from the DAO enter the verification process. In this stage, the selected verifiers from the DAO use AI tools and other methods to verify the technical quality, artistic value, theme of the artist’s representative artwork, and check if there are any copyright issues with the artist’s creations. The artists and creators who pass the verification are registered, and the verifiers receive compensation from the DAO. In the third step, the selected artists and base image creators undergo evaluation for the artwork they registered with AIU. After registering the artwork with AIU, it goes through a verification process by various general evaluators for a certain period. Anyone can be a general evaluator and evaluate the artwork. General evaluators investigate whether there are any copyright issues with the artwork submitted by the artist, and if there are any issues, the corresponding image is deleted. In this stage, evaluators utilize their experience and expertise to perform verification, and artists who succeed in verification can continue their activities with AIU. (If an artist applies using AI generated artwork, they will face sanctions during the verification and evaluation process.)

2. Training

AIU uses various methods to train AI using the artwork provided by the original creators. One of the main methods is called fine-tuning, which is a modeling technique that fine-tunes model parameters by training additional datasets based on a pre-trained model. Fine-tuning allows for more efficient and faster results by adding new datasets to an existing model compared to training a model from scratch.

3. Generating

1) Base image selection process The base image selection process is the process of selecting secondary products that are closer to the desired image for generating compared to when users generate images. In this process, users select the desired base image from the registered base images in AIU. The selected base image is used for image generation, improving the completeness and accuracy of the secondary products.

2) Original image selection process AIU provides users with the option to choose the desired art style from the artists' artwork. Through this, users can create their own creations by selecting various art styles. The process of selecting the artist's art style is as follows: First, users look through various artists' artworks provided by AIU to find the desired style. Users can view thumbnails or previews of the artworks provided by each artist. Next, users select the artist with the desired style. Users can check and select based on the artist's name or artwork title. If a user wants to combine the art styles of two artists, AIU provides a feature to combine the art styles of the selected artists to create a new style. This allows users to combine different styles to create their own artwork.

3) Image output using AI generator When the user selects the desired art style, the AI generator creates a new image following that style. Through this, users can obtain an image with the desired art style. The number of images provided by the AI generator to the user varies depending on the model used, parameters, input image size, and diversity of generated images. Generally, multiple images are generated for users to choose from. Furthermore, an image enhancement process is conducted to make the selected image closer to the desired image. This process analyzes the generated image to find the most similar image to the desired style with further enhancements using that similar image. Finally, the AI generator saves and delivers the image the user desires. This allows users to generate images of various styles with a few simple touches.

4. Issuance and Trading of NFTs

AIU users can publish their desired images as NFTs after generating the images using AIU. By publishing an image as an NFT, the user acquires ownership and on-chain copyright (referring to copyright on-chain, not legal copyright) of the image. When an image is published as an NFT, ownership and information about the copyright holder are recorded on the blockchain. While ownership can be transferred if the user moves the NFT to a different wallet, the copyright remains unchanged, and the user may continue to receive ongoing royalties each time the NFT is traded. If the transaction is made in a marketplace that supports copyright fees, by generating and publishing NFTs based on works that the artist has agreed to utilize, users can have more liberty regarding copyright issues by generating and publishing NFTs.

5. Reward System

AIU operates with the participation of various members such as base image creators, artists, validators, and general evaluators. Each member receives AIU token rewards for contributing to the services provided by AIU. Base image creators and artists can provide the necessary data for image generation and receive rewards. Validators and general evaluators verify aspects such as image authenticity to contribute to the development and operation of AIU and receive appropriate rewards from the DAO.