, an AI program developed by startup Stability AI, which generates images based on text, was open sourced, so someone soon started to use this AI to generate color images. But early AI-generated works are unsightly and are simply a distorted one.
When everyone thought some things were still irreplaceable, they didn't expect that they were slapped in the face in just a few months. In the past few days, some pictures of unknown origin have been reposted crazily, and it is not an exaggeration to praise the quality of the work of art. But things did not end there, but continued to spread to other areas.
astringent picture? It is clearly exploring the art of the future!
Since Stable Diffusion, the AI technology has been open sourced, users of a certain forum have begun to use the AI to generate images of real and animated nude characters, most of which are women. Of course, there are some strange pictures of ...
Although the forum manager quickly closed many sections dedicated to artificial intelligence sluggish pictures, these people began to set up their own businesses. Until now, the largest Unstable Diffusion is building business around AI, is built to generate higher quality maps and currently earns more than $2,500 per month from hundreds of members to maintain community operations.
Now, Unstable Diffusion can use AI technology to generate a series of pornographic pictures of different artistic styles and divide them into different channels. Currently, they have generated more than 4,375,000 images, and the organization also hosts competitions from time to time to challenge members to reconstruct images using robots and then in turn to improve the model.
With the development of Unstable Diffusion, it aspires to become a "moral" community for artificial intelligence to generate pornography—that is, a community that prohibits child pornography, deep fakes and excessive blood. However, the server's users must comply with the terms of service and be reviewed for the images they generate. Administrators claim that the server used a filter to block images of people with their "celebrity" database and had a full-time audit team.
Such a rich style naturally attracts a large number of users to participate. During the discussion on Twitter , everyone is sharing their latest "masterpieces" every day, and everyone on the Internet applauds for this platform. People say this is exploring the future of artificial intelligence!
grasps the tail of technology, and supervision is no problem.
Before, there was a tool for AI-based novels Novel AI. In order to facilitate the author to provide pictures of the article, the team developed an AI drawing function, but the painting became popular.
Novel AI even supports saving seeds based on the output results for convenience in future use. If you want to "create"
Novel AI also has a magic pen Ma Liang function. As long as you have hands, you can become a "big touch". You can convert a few strokes and one click to convert it into detailed illustrations.
, which was originally limited to the two-dimensional field, has begun to shift the attention of many creators as the World Cup begins. They began to use this AI to to synthesize some celebrity players’ appearance of stripping naked!
Of course, Novel AI needs to consume points to generate images, so as long as you spend money to recharge, you can get fixed points every month. The prices are $15/month, $20/month, and $25/month. There are three levels to choose from. The more expensive the points, the more points you get. The highest-end generator has the strongest generation capabilities!
So how should we supervise such a powerful AI? The past lessons of the past have caused quite a bad impact, and AIs like Deepfake are used to do evil are very harmful. Recently, Intel developed a detection tool called FakeCatcher, and claims to be 96% accurate in detecting videos tampered with by AI.
The principle is that judges by detecting blood flow. As blood flows through our body, our blood vessels changes color in very subtle ways, which is difficult for the human eye to detect, but can be detected by optical density measurement.
So Sofia thinks that the best way to regulate AI is to keep up with technology and use technology to deal with technology. Only in this way can AI be firmly tied up and benefit society rather than harming society.
