Mind-controlled typing: Meta develops AI system to translate brain activity into text

CALIFORNIA: Meta’s groundbreaking discovery has made the vision of mind typing a reality, somewhat. The AI and neuroscience team has guided the deep learning system to read magnetic signals sent out by the brain to find out which keys any person has mentally selected, without using implanted electrodes or wearable sensors.

Hearing those terms might elicit shock when spelled with magnetoencephalography (MEG), otherwise known as the finer, more correct term for anything that captures brain activity through magnetic fields.

Volunteers are seated within a MEG scanner, a large contraption roughly the size and shape of a giant hair dryer, allowing them to type a sentence in addition to recording magnetic field signals from neurons.

Brain2QWERTY is a model that the use of training teaches how magnetic signals correlate to different letters so it can work out and reconstruct certain words with up to 80 percent accuracy.

This does, however, raise challenges that should be concluded with caution before furthering curved repairs. At this point, we’re looking at a heavy machine that weighs half a ton and includes a $2 million price tag that’s far from portable. Users have to remain completely still, with any minor movements of the head disrupting the process.

The MEG scanner must also be shielded from interference from Earth’s magnetic field, one trillion times stronger than the signals from the brain. Slight movements, also, scramble the data, and that makes it hard to put consistent results forth.

Nevertheless, the risky technology might be your saving grace, presenting new hope for patients who suffer from brain injury and neurological conditions, as well as paralysis, to communicate once again.

Unlike Neuralink’s invasive brain implants that attempt to tie human thought into computers, Meta’s technology is non-invasive, which avoids some of the surgical risks that could pose an obstacle to long-term success and makes the technology much more palatable.

Although current technologies are only at an early stage, they may lead to breakthrough improvements and into one of the holy grails in understanding.