While everyone is discussing ChatGPT, another powerful AI software, Stable Diffusion, has slowly started to take away many artists’ jobs. This AI drawing software, named “Stable Diffusion,” immediately beat a host of competitors upon its release, becoming not only a great assistant for artists but even taking away their own jobs.
With just a powerful graphics card PC, a few keywords, and a couple of model downloads, you can generate artworks that would take a comic artist many hours to create:
For example, these three comic images require a lot of time for artists to outline details, but ordinary people, after downloading the exclusive models, only need to input a few keywords:
masterpiece, best quality, intricate detail, artbook, ultra-detailed, illustration, ray tracing, close-up, (full body), wearing thigh boots, standing, dress, straight on, face focus,
The images are generated, with requests for styles of master paintings, best quality, complex compositions, ray tracing, high boots, standing…
If you want to merge real-life figures with your creative ideas, it’s also very simple. For example, the star Emma Watson is portrayed as a powerful sorceress casting lightning magic:
Add a few keywords, and Emma Watson is a powerful sorceress casting lightning magic, surrealism, full body painting:
Similarly, these keywords, combined with the style of a specific artist, can produce different results. For example, this time the artists are Stanley Artgerm Lau and Alphonse Mucha:
And so on… A simple search online reveals many such creative ideas.
For instance, Tifa in the style of Akira Toriyama:
Aside from fingers not being rendered correctly, there are no significant flaws.
It can also generate such refined images:
This leaves influencers who take hours to edit photos with nowhere to hide; it can be generated in just one minute.
Many industry leaders believe that Stable Diffusion is an open-source bomb. While OpenAI has ample funding, ChatGPT is difficult to compete with it from a commercialization and product perspective.
Emad Mostaque, founder of Stability AI, believes that images are the killer application. Image generation models can quickly create and guide people to consume rapidly.
Stability AI, which supports investment in Stable Diffusion, is currently seeking funding with a valuation of $4 billion. Investors seeing such results will likely flood it with money.
Of course, such a powerful Stable Diffusion also raises some concerns, such as whether generating images of Emma Watson infringes copyright? In the future, as more learning models emerge, will generating images in this way, even creating fakes, be somewhat unethical? After learning artists’ styles, do the subsequently generated images infringe copyright? For ordinary artists, before making a name for themselves, has their foundational work already been taken by AI?