Hajra Naz
Meta is testing a device that translates brain activity into on-screen text — but only if you’re ready to haul around a half-ton piece of lab equipment.
The new system can accurately identify about 80% of the letters you “type” with your mind, yet it remains far from practical.
The setup costs around $2 million, requires a specially shielded room to block out Earth’s overpowering magnetic field, and even a slight head movement can throw off its accuracy.
While the idea of mind-to-text might sound like a scene out of science fiction, it’s clear that this isn’t a tool for everyday use.
Imagine lugging a full-size MEG scanner, roughly the size of a refrigerator, just to compose an Instagram caption. This is more of a laboratory experiment than a portable gadget, at least for now.
Meta’s breakthrough comes from combining artificial intelligence with advanced neuroscience.
Their teams have trained a deep neural network to decode external brain signals and map them to specific keystrokes — all without implants or futuristic headbands.
Read More: Meta unveils new program to improve speech and translation AI
In a pair of new studies, the research details how, after extensive training, the system can reconstruct full sentences from thought alone.
In the experiment, participants sit inside a MEG scanner that resembles a giant hairdryer.
The scanner captures magnetic signals generated by neuronal activity, and an AI model — dubbed Brain2Qwerty — learns to associate these signals with corresponding keys.
Although the results aren’t perfect, achieving up to 80% accuracy is a promising step toward understanding how our brains convert thoughts into written language.
The technology still has its hurdles. Beyond the impracticality of the heavy, expensive equipment, even tiny head movements can distort the signal.
Yet, this research might have far-reaching implications, not only for advancing our understanding of brain functions but also for developing new treatments for brain injuries and neurological disorders.
Meta explained in a recent blog post,
“By capturing 1,000 brain snapshots per second, we can observe the precise moment when abstract ideas morph into concrete actions like typing.”
Their study reveals that the brain shifts from a high-level concept of meaning to the fine motor actions required for typing—a process that might one day lead to wearable brain-interface devices.
