Meta invents device that replaces a keyboard for typing with thoughts

Meta has invented a device that is capable of replacing a keyboard for typing with a user instead thinking of the text they want to produce.

Two studies published by Meta last week detail researchers creating a new state-of-the-art brain scanner and combining it with a deep learning AI model that's designed to interpret brain signals and convert them into text. Here's how it works. The new platform consists of a magnetoencephalography scanner, which is a neuroimaging technique designed to measure and record the magnetic activity produced by the electric currents generated by brain signals.

The activity recorded by the machine is then combined with the deep learning AI model that interprets the signals into the keys pressed by an individual. The AI model eventually reached a level of accuracy that entire sentences could be reconstructed. Researchers said the system was capable of correctly predicting the keys typed by a "skilled" typist 80% of the time. Meta says the new system is the most accurate brain-typing system that doesn't involve an invasive procedure such as a brain-computer-interface (BCI), which involves surgery and connecting a device to the brain.

However, Meta's new platform has quite notable limitations, such as the brain scanner costing $2 million, weighing more than a ton, the user being tested being required to keep their head perfectly still for accurate readings, and the room the scanner is placed in to have special dampening walls to block out Earth's magnetic field as it would contaminate the test results with rogue magnetic readings.

All of the aforementioned factors lead to the assumption that this technology isn't likely to be released commercially. However, it's application may unlock new levels of artificial intelligence.

"Trying to understand the precise architecture or principles of the human brain could be a way to inform the development of machine intelligence," Jean-Rémi King, leader of Meta's Brain & AI team, told Tech Review. "That's the path."

"Language has become a foundation of AI," King told MIT. "So the computational principles that allow the brain, or any system, to acquire such ability is the key motivation behind this work."