(FiveNation.com)- New mind-reading technology has been revealed at the University of Osaka, Japan, according to The National Pulse. Researchers used a combination of artificial intelligence and magnetic resonance imaging (MRI) to form images that come from the human brain. The effort was reportedly low cost as they were able to use “Stable Diffusion” technology, which converts text into high-resolution images.
The study concluded that the image reconstruction process can be done without any training or other deep-learning models. This is different than the ChatGPT model which is based on fine-tuning and a large language model (LLMs). Deep-learning technology to recreate images from the mind has usually been required up to this point before researchers began to use linguistic prompts.
The possibilities for this new technology have already started to cause some artists to believe they will be replaced. Efforts to get the technology restricted or banned are underway. The winner of the September 2022 Colorado State Art Fair competition, Jason M. Allen reportedly said that “Art is dead,” after his AI-generated “Théâtre D’opéra Spatial” took the blue ribbon.
The mind-reading technology requires a two-step process. First, the researchers reportedly decoded the visual images from the MRI and then, secondly, they turned those codes into text representations which were put into Stable Diffusion to generate a higher-quality version of the image.
The system used by the researchers is also used by millions around the world already. Stable Diffusion reportedly trains itself using the images that it has been fed. While it might be considered a search engine, it displays images that do not actually exist.
Artificial intelligence has potential that is cause for worry, according to some critics. With ChatGPT, students are able to generate complicated essays by inserting a prompt. Students are already reportedly using the system for their work.
