Artificial intelligence (AI) is now able to interpret human thoughts and create AI-generated images. This feat was accomplished by researchers from Stanford University, the National University of Singapore, and the Chinese University of Hong Kong. They designed technology to reconstruct images from recorded human brain activity. While the images created were not perfect, the team was able to achieve an 84% success rate in generating images that matched the original photos’ attributes, such as color and texture. The technology is focused on understanding human brain activity and using it to decipher what people are thinking. It could be useful in fields such as medicine, psychology, and neuroscience, offering potential treatments for neurological disorders and allowing nonverbal communication through thoughts.
The Experiment and Its Process
The team collected data from participants’ brain activity while they viewed pictures. Participants’ thoughts were then recorded inside a functional magnetic resonance imaging (fMRI) machine. An AI model, called Mind-Vis, was trained on a pre-existing dataset of over 160,000 brain scans. The AI was then trained on a small dataset from participants’ brain activity recorded by an fMRI machine as they viewed images. The AI learned to link specific brain activities with visualizing image features such as color, shape, and texture. Each participant’s brain activity was measured for 20 hours during training. The AI was then able to generate images based on assessments it made from participants’ brain activity.
Several similar brain activity and AI-image experiments have been conducted recently. In an experiment by researchers from Japan’s Osaka University, participants viewed images while their brain activity was recorded. However, their study did not involve training an AI model to create the images. In another experiment by a Russian team, an AI sketched what a person wearing an electrode cap was seeing at that moment. While the technology still has a way to go before it can be widely used, it has promising applications in medicine, psychology, and neuroscience. As the technology continues to improve, it could potentially be used to allow people to control a computer by just thinking of the commands instead of typing on a keyboard. In a perfect world, people could even send messages just by thinking.
Leave a Reply