In recent months, the trend of artificial intelligence (AI) has taken the social media world by storm, leading to the emergence of various types of AI through the Internet. It is evident that AI is rapidly evolving, developing strongly, and being utilized in many aspects of life.
This phenomenon is largely attributed to the excellent “learning methods” of AI, which help it generate outstanding results that users desire in a short time. As more people begin to use AI, it will increasingly refine its outputs to better match expectations and provide more relevant visual representations compared to human-generated content. However, it is important to note that these AIs operate based on a model of a text chain that they transform into visual representations.
For example, if a user inputs: “Michael Jackson looks like a character in the anime JoJo’s Bizarre Adventure“, the AI will process the specific parts of the input and return image results. Of course, the resulting output depends not only on the AI program but also on the user’s input; if the user provides a very broad prompt, the results might not meet their expectations.
Moreover, a user from Japan utilized LORA (one of the prevalent AI models available on the Internet today) and asked the AI to transform an image of Yamcha, a character from Dragon Ball Z, into a loli character.
As a result of this request, many users have expressed interest, leading to a plethora of AI-generated images depicting Yamcha as different characters from various anime, which have been widely shared on social media. However, among the shared images, some still require significant improvements in quality. They often suffer from fundamental errors that prevent viewers from fully appreciating the artwork.
