Stable Diffusion is an image generator that can create what looks like real photographs or hand-crafted illustrations. The properties of a large collection of images taken from the web and image databases can be associate with their associated text labels thanks to a program. Random noise is added to an image to make it match a prompt.
Stable Diffusion's software is capable of generating new sexually explicit pictures due to the fact that their training data includes pornographic images. One concern is that the images could be used to spread misinformation.
In the past year and a half, the quality of artificial intelligence-generated imagery has increased. The model of generating images from text prompt was popularized and followed by a more powerful successor, DALL-E 2.
Access to the image generators has been restricted since the beginning, with only a prompt that filters what can be requested allowed. A service called Midjourney was released in July of this year that helped popularize the use of artificial intelligence in art.
Stable Diffusion is not the first of its kind. The DALL-E Mini was created by a developer after the original DALL-E was released and quickly became a meme. The Mini version of DALL-E has the same guardrails as the official version. Clement Delangue, CEO of HuggingFace, a company that hosts many open source artificial intelligence projects, says it would be difficult for the technology to be controlled by a few large companies.
He says that if you look at the long-term development of the technology, it's actually better from a safety perspective. It is better if outsiders can assess models for problems such as race, gender, or age biases in order to understand closed technology. The benefits of open source technology outweigh the risks according to him.
Delangue suggests that social media companies could use Stable Diffusion to build their own tools to spot fake news. Adding invisible watermarks to images made using Stable Diffusion so they are easier to trace, and building a tool for finding problematic images in the model's training data are some of the things developers have contributed to.
Simpson- Edin became a moderator on the Unstable Diffusion Discord after becoming interested in Unstable Diffusion. People are not allowed to post images that could be seen as pornography. She says that they can't moderate what people do on their own machines. Artificial intelligence may affect humans more than machines in the near term.
All Rights Reserved © 2024
