Three Years of Daily BCI Use Proves Neural Interfaces Can Leave the Lab
Casey Harrell's sustained ALS speech implant use shifts brain-computer interfaces from demo milestone to practicality benchmark, pressuring Neuralink and rivals.
8. Three Years of Daily BCI Use Proves Neural Interfaces Can Leave the Lab
Casey Harrell, a man living with ALS-related paralysis, has used a brain-computer interface to generate speech for nearly three years, clocking thousands of hours on the device since a research team first activated it in 2023. The implant, a set of electrodes embedded in his motor cortex, decodes intended speech signals and translates them into audible output. Harrell is now described by the research team as "the first power user" of the technology, a framing that signals something specific: this is no longer a proof-of-concept measured in minutes. It is a system someone depends on daily.
That framing matters competitively. Neuralink, which implanted its first human patient in January 2024 and demonstrated cursor control and limited communication, has consistently led the public narrative around BCI practicality. Harrell's case, developed through academic research rather than a venture-backed company, now sets a durability benchmark Neuralink has not yet publicly matched for speech specifically. Synchron, which uses a less invasive stent-based electrode approach and has its own ALS patients, faces the same comparison: thousands of hours of sustained use is a harder target than a highlight reel. For the BCI field broadly, the question shifts from "can it work?" to "can it hold signal fidelity and user trust across years?" That is a different engineering and regulatory problem.
The pattern worth watching is how academic BCI programs and commercial players respond to a sustained-use standard. FDA clearance pathways for implantable BCIs currently treat long-term biocompatibility as a key gate. Multi-year real-world data from non-commercial research implants may accelerate or complicate those reviews, depending on what the signal degradation curves look like. If Harrell's team publishes longitudinal performance data, it becomes a reference dataset every commercial BCI applicant will need to address.
Source: MIT Technology Review