AI image generation
이미지 생성 AI
AI that creates a new image, such as a picture or a photo-like image, from a description in words.
Overview
Write something like "a puppy walking on a beach at sunset," and a picture appears in a few seconds. This works because the AI learned from huge numbers of pictures paired with their descriptions. The diffusion model method, which starts from noise and polishes it step by step into a picture, is widely used.
Besides drawing new images, it can also change the background of a photo, erase objects and combine several photos. It is widely used in advertising, design and games to produce rough drafts fast. It can also easily make fakes that look like real photos, so the more surprising a photo is, the more you should check its source.
An everyday example
It is like ordering a painting out loud: "please paint me this picture." The difference is that it paints several in a few seconds, and repaints as many times as you like if you are not happy.
Why it matters
Anyone can now turn a scene in their head into a picture, even without drawing skills. At the same time, disputes over copyright and trust have grown, over imitating an artist's style and over fake photos.
Also calledimage generation, image generation model