The hat reads the wearer's thoughts and types them into text

Mind reading may soon become a reality, after researchers at the GrapheneX-UTS Artificial Intelligence Centre at the University of Technology Sydney (UTS - Australia) were able to convert thoughts in the brain into text on a computer screen.

The research was selected for publication at the NeurIPS conference, an annual meeting of researchers in artificial intelligence and machine learning.

At the conference, test participants silently read passages of text, while an AI model called DeWave — using only their brainwaves as input — projected those passages onto a screen.

With this new technology, users only need to wear a hat that can record their brain activity through electroencephalography (EEG), which is much more practical and convenient than previous invasive technologies.

The researchers admit that with this initial study, the signal was a bit noisier than information obtained from a brain implant. Additionally, the model was more adept at conjugating verbs than nouns.

When it comes to nouns, we found that the model tends to use synonym pairs rather than direct translations. For example, the machine translates “man” instead of “author,” explains Yiqun Duan, the study’s first author.

“We realized that semantically similar words can produce similar brain wave patterns,” said Duan.

However, researchers believe they can improve this accuracy.

“We will continue to align keywords and similar sentence structures, hoping this model will bring meaningful results to more patients with speech impairment,” said Duan.

Ching-Ten Lin, director of the GrapheneX-UTS Center, said the research "marks a significant breakthrough in the field of BrainGPT", a pioneering effort to translate raw electroencephalogram (EEG) waves directly into language.

“For the first time, researchers have combined discrete encoding techniques in the brain-to-text translation process, introducing an innovative approach to neural decoding. Integrating large language models is also opening new frontiers in neuroscience and AI,” he said.