As artificial intelligence continues to evolve, it shows off its potential to outperform humans in various tasks and domains. Now, scientists achieved a milestone with their study regarding mind-reading AI that transforms thoughts into written text.
A team of scientists at the University of Technology Sydney (UTS), Australia, developed a groundbreaking technology utilizing a sensor-covered helmet. This technology examines specific electrical activity in the brain as the wearer thinks, translating these signals into spoken words.
The recorded electroencephalogram (EEG) signals were transformed into text using an AI model known as DeWave. In a demonstration video below, a person is seen contemplating a sentence displayed on a screen.
Mind-Reading AI: Sydney Tech Researchers Unveil a Breakthrough
Researchers from the University of Technology Sydney have unveiled what seems like a leap into the future: an AI that translates thoughts into text.
The system reads EEG signals and converts them into words and… pic.twitter.com/tQ0lFQKwOZ
— Bad AI (@Bad_AI_) December 12, 2023
Rising of mind-reading AI
According to Chin-Teng Lin from the UTS, the technology is non-invasive, relatively affordable, and portable, reported New Scientist. “This research represents a pioneering effort in translating raw EEG waves directly into language, marking a significant breakthrough in the field,” he said.
This new technology is a giant leap forward because it uses special techniques to turn brain activity into text, and this introduces a fresh way of understanding and decoding how our brains work, says Chin-Teng.
The system is not flawless, with an accuracy of about 40 percent. Newer data is currently under peer review and indicates an enhanced accuracy surpassing 60 percent.
The DeWave model underwent training by examining numerous instances. “For instance, when you think about saying ‘hello,’ your brain sends out certain signals,” explains Ching-Theng. DeWave understands the connection between specific brain signals and the word ‘hello’ by studying numerous instances where these signals occur in different words or sentences.
Credit: University of Technology Sydney
After mastering the understanding of brain signals, the team connected DeWave to an open-source large language model (LLM), resembling the AI that drives ChatGPT. The team carried out collaborative training for both DeWave and the language model to improve their proficiency in generating sentences using the EEG data.
Elon Musk’s Neuralink
Meanwhile, Elon Musk is known for his efforts to make a direct connection between human brains and machines. Musk, in this regard, established Neuralink in 2016 intending to develop brain–machine interface (BMI) technologies.
Neuralink aims to create devices that can be implanted into the human brain, allowing for direct communication between the brain and external devices, such as computers.
