Neuralink Demonstrates Mind-Controlled Robotic Arm 'Steel Kiss'; Advances Toward Full Digital Control via Dual-Implant System

Elon Musk’s Brain-Computer Interface: New Developments

In Neuralink’s official video, a demonstrator has achieved mind-controlled operation of a robotic arm.

The robotic arm steadily lifted, extended toward the presenter’s mouth, and delivered a kiss.

Musk also stated that, in theory, Neuralink will be able to indirectly control everything via computers or mobile phones.

Meanwhile, Neuralink is advancing an upgraded trial.

First Subject May Receive Update

Neuralink’s Upgrade Trial to Adopt “Dual-Implant” Approach; First Human Subject Noland Arbaugh Expected to Be the First Patient to Undergo a Second Surgery

The plan involves adding a second interface at the spinal cord location while retaining his existing cranial implant.

The purpose of this architecture is to establish a “digital neural bridge,” where the chip at the brain end decodes motor intentions, and the chip at the spinal cord end receives commands and directly stimulates limb nerves, thereby bypassing damaged biological pathways in an attempt to restore walking ability in paralyzed patients.

In terms of hardware stability, the technical team has completed key iterations to address the "electrode wire retraction" issue exposed in the first-generation product.

In response to the first patient’s electrode lead dislodgement and signal attenuation caused by intracranial air pockets, the team refined the surgical protocol for the second subject, Alex, and subsequent patients by strictly controlling the gap between the implant and the brain surface (Gap Reduction) and optimizing intraoperative strategies to prevent brain tissue displacement.

These measures have effectively eliminated the phenomenon of physical retraction, ensuring long-term signal stability. The current R&D roadmap indicates that the technological focus is shifting from unidirectional EEG reading to complex bidirectional “brain-spinal cord” signal bridging.

Musk revealed that this multi-implant synergistic architecture is not only intended for medical restoration in the future, but also plans to significantly increase information transmission bandwidth through hardware stacking, with the ultimate goal of enabling subjects’ reaction speeds and digital interaction efficiency to surpass normal human levels.

New Uses for Brain-Computer Interfaces: User-Led Exploration

In addition to the officially designated functions, Neuralink’s subjects are also actively exploring more applications of brain-computer interfaces in daily life.

Brad Smith, Neuralink’s third human subject, is a typical case.

Although he had become proficient at controlling the computer cursor with his thoughts, he quickly realized that the cursor alone could not overcome the physical visual field limitations caused by ALS.

Specifically, due to his inability to rotate his neck, his gaze was locked onto the computer screen directly in front of him.

Smith did not wait for the medical team to develop a dedicated visual aid; instead, he conducted his own research online and selected a standard conference webcam.

This request initially puzzled the Neuralink team, which focuses on neural signal transmission, but under Smith’s insistence, engineers assisted him in mounting the consumer electronic device onto his wheelchair and integrating it with the control system.

Now, Smith can simply imagine hand movements in his mind to click on the control panel on the screen and direct the camera to rotate and zoom.

To accommodate his usage habits, the camera manufacturer even specially adjusted the software parameters, enabling a single thought-controlled click to drive the lens to rapidly pan at three times the normal speed.

With this makeshift system, he was finally able to zoom in to clearly see the wedding photo on the wall and observe the expressions of guests at the table during family gatherings via the screen.

This user-led exploration is not an isolated case in the brain-computer interface community.

Nathan Copeland, a veteran user who has had another brain-computer interface implanted for over a decade, also stated bluntly that the brain-computer interface is merely a tool in his toolkit, which must be used in conjunction with ordinary devices in daily life.

To control the lights and television in his home, he directly uses the Google Home voice assistant; for playing video games, he even sought help on forums to have arcade joystick and Xbox controller components soldered together, creating a specialized controller tailored to his operational needs.

These users did not confine themselves to the role of medical test subjects; instead, by connecting expensive implants with affordable cameras and smart home devices, they regained their autonomy in observing the world and controlling their lives.

Neuralink in the Past Two Years

2024–2025: Neuralink Achieved Its First Human Implants and Made Substantial Progress Across Multiple Technical Fronts

In January 2024, Neuralink successfully implanted its N1 brain-computer interface chip into a patient with high-level paralysis for the first time.

The surgery was performed by Neuralink’s self-developed R1 robot. The chip interfaces with the brain’s motor cortex via 1,024 flexible electrodes, enabling the postoperative system to acquire neural signals in real time and wirelessly transmit them to external devices via Bluetooth.

A few weeks later, the patient successfully controlled a computer mouse and operated a typing interface using thought alone during a live broadcast, demonstrating that the closed-loop brain-computer interface system had achieved initial functionality in a human subject.

In the following months, Neuralink continued to optimize signal decoding accuracy, device battery life, and safety, and began recruiting additional participants.

In September 2024, its “Blindsight” project for visual restoration received the U.S. Food and Drug Administration (FDA) Breakthrough Device designation, aiming to enable basic visual perception in blind users by stimulating the visual cortex.

In May this year, its system for auxiliary language functions also received the same certification, with plans to develop a “brain signal → speech/text” pathway for patients with aphasia.

By this September, Neuralink announced that it had completed implantation surgeries for 12 subjects. The cumulative device operation has exceeded 15,000 hours, with no major rejection reactions, and overall stable performance.

This marks Neuralink’s transition from single-case validation to multi-case deployment. Meanwhile, the company stated that it is preparing for additional international clinical trials and has already obtained ethical and regulatory approvals in countries such as the United Kingdom and Canada.

Over the past two years, Neuralink has achieved the first human implantation of its device, validated mind-control functionality, expanded clinical trials to multiple cases, and obtained multiple FDA approvals, thereby establishing the practical feasibility of its early applications in the field of brain-computer interfaces.

In the next phase, its focus will gradually expand from motor control to more complex neural functions, such as language and vision.

One more thing

However, what has recently gained even more widespread attention regarding Neuralink is the latest revelations from Musk’s interview.

As the interviewer had an India-related background, Musk voluntarily brought up Shivon Zilis, referring to her as his “partner,” and mentioned that one of the children they had together was given “India” as a middle name.

Shivon Zilis, 38, an Indian-Canadian, is an executive at Elon Musk’s Neuralink.

Reference Links:

[1]https://x.com/elonmusk/status/1995945912510939622

[2]https://www.teslarati.com/neuralink-first-patient-could-receive-an-upgrade-elon-musk/

[3]https://www.theverge.com/report/829120/neuralink-bci-webcam

This article is from the WeChat official account: QbitAI (ID: QbitAI), by Kresy