Max Headroom 
Neuralink, Elon Musk’s brain-computer interface company, is reportedly planning to start a US clinical trial in October to use its technology to translate thoughts directly into text.
The trial was approved by the FDA under an investigational device exemption, and it could help pave the way for people with speech impairments to communicate more quickly and efficiently.
Thought-to-text technology
The upcoming trial aims to read speech directly from the brain, enabling participants to transmit words without moving their mouths or using a keyboard, as noted in a report from Bloomberg News.
DJ Seo, Neuralink’s president, told an audience at the Korea Foundation for Advanced Studies in Seoul that the company hopes to demonstrate communication with large language model AI platforms “at the speed of thought,” potentially faster than spoken speech.
“We think that it’s actually possible to demonstrate abilities to speak to the latest AI model, or LLM models, at the speed of thought, even faster than how you’re speaking, and being able to potentially get that information back through your AirPods, effectively closing the loop,” Seo stated.
Production ambitions
Ultimately, Seo described the trial as a stepping stone toward broader human enhancement and consumer-facing applications, beyond the company’s initial focus on medical conditions like ALS, stroke, blindness, and Parkinson’s disease.
While initial implants will focus on patients with severe speech impairments, Neuralink plans to eventually expand to healthy individuals by 2030. The company aims to scale to 20,000 implants per year by 2031.
“We’re currently envisioning a world where, in about 3 to 4 years, there will be someone who’s otherwise healthy who’s going to get a Neuralink. If you’re imagining saying something, we would be able to pick that up,” the executive noted.
The post Neuralink is planning to launch US trials translating thoughts into text appeared first on TESLARATI.
Elon Musk, News, Featured, Neuralink
TESLARATI
[crypto-donation-box type=”tabular” show-coin=”all”]
