Compiled by the editorial desk with reference to the study published in Cell and reporting from the New York Times.

Researchers at Stanford University have developed a brain-computer interface (BCI) that can translate unspoken words into audible speech for people who have lost the ability to speak, according to a study published in the journal Cell. The technology, tested on four volunteers with conditions like amyotrophic lateral sclerosis (ALS), aims to reduce the physical strain of communication by decoding neural signals associated with internal speech.

Erin Kunz, a Stanford neuroscientist and coauthor of the study, told the New York Times that the motivation came from observing how exhausting traditional BCI use can be. Standard systems require users to attempt vocalization, which can be tiring for those with weakened airway muscles. The new approach seeks to bypass that effort by reading thoughts directly.

“If we could decode that, then that could bypass the physical effort,” Kunz said. “It would be less tiring, so they could use the system for longer.”

The study builds on earlier work with Casey Harrell, an ALS patient who had already used a BCI to speak through brainwaves and recorded podcast interviews from before his condition progressed. In the new phase, researchers found that decoding internal speech was more challenging than expected, so they retrained their AI models to better match thoughts to words. The system eventually translated complex sentences like “I don’t know how long you’ve been here” with improved accuracy.

Privacy safeguard emerges

During the trial, the team noticed an unexpected issue: the computer sometimes picked up words that participants were not intentionally thinking, raising concerns about mental privacy. To address this, they introduced a mental “password” that users imagine before and after a phrase they want spoken. The chosen word was “Chitty Chitty Bang Bang,” a reference to Ian Fleming’s 1964 novel, chosen for its distinctiveness to avoid accidental triggers.

When participants used this inner password, the system correctly activated and deactivated decoding 98.75% of the time, according to the study. Kunz described the trial as a “proof-of-concept,” but the results suggest a path toward more practical and private BCI use.

The research highlights both the promise and the ethical challenges of brain-reading technology. While the system could offer a less exhausting way for people with speech impairments to communicate, it also raises questions about which thoughts should remain private. The Stanford team’s approach—using a user-controlled toggle—offers a potential solution, though larger studies are needed to confirm its reliability.

As BCI technology advances, the balance between functionality and privacy will remain a critical consideration. For now, the Stanford study provides a glimpse of a future where unspoken words can be shared at will, without broadcasting every stray thought.