Introduction to AIGC
What is AIGC?
AIGC is a technology that generates content using artificial intelligence models.
The content generated by AIGC can include images, audio, text, video, 3D models, and more.
If we break it down further, this technology can generate the following types of content:
- Images
- Photos
- Original artwork
- Audio
- Voice (e.g. voice acting in a video game)
- Music (e.g. Taylor Swift songs)
- Text
- Code (e.g. a block of C++ code)
- Copywriting (e.g. ad copy)
- Novels (e.g. fan fiction)
- ...
- 3D Models
- Characters (e.g. Genshin Impact waifus)
- Scenes (e.g. buildings)
AIGC Product Forms
For the average user, AIGC technology is typically used to generate a specific type of content, such as creating images of beautiful women, generating a voice that sounds like a particular voice actor, continuing a novel, or creating music in a specific style.
Currently, in the early stages of AIGC development, the most common product forms are text-based and controlled through user input. The typical product interaction is:
Input: Text description of desired content
Output: Content generated from the text description (image, music, text, video, etc.)
Here are some examples of content generated by AIGC:
Text-to-Image Generation
Now, let's say I want to generate a beautiful landscape painting:
Input (Text Description): A landscape painting with a tall mountain peak in the center, shrouded in clouds.
Output:
Text-to-3D Model Generation
Now, let's say I want an airplane that looks like a banana:
Input: An airplane that looks like a banana.
Output:
Text-to-Text Generation
In addition to text-based interactions for generating other types of content, there are other product interaction forms, such as:
Image + Text-to-Image Generation
Input:
- Text: Make this penguin look like a cartoon.
- Image:
Output:
Image-to-3D Model Generation
Input:
Output:
Overall, the current form of AIGC products is user input of a text description of desired content, with the model outputting the desired content.