Brain-Computer Interfaces (BCI): Towards Humans’ Merger with AI 

The ongoing clinical trials of Brain-Computer Interfaces (BCIs) such as Neuralink’s “Telepathy” implant involve establishing communication links between the brains of participants who have unmet medical needs due to damaged biological interfaces in conditions such as spinal injury, stroke, or amyotrophic lateral sclerosis (ALS) and AI platforms. The BCI implant takes over the function of the damaged biological interfaces, and the participants of the trial are able to use phones, computers, laptops, games, and robotic arms by thought alone. This progress indicates that, in the near future, it may become possible to establish a high-speed connection between the brain and AI platforms bypassing our abysmally slow biological interfaces and overcoming bandwidth constraints to integrate AI into our tertiary compute level. The high-bandwidth neural links would serve as a bridge, effectively merging the brain with AI. Humans would become cyborgs (cybernetic organisms). The merger would enable the two to benefit from each other. The brain would acquire superhuman computing power of AI thereby mitigating the risk of humans being obsolete in the face of superintelligent digital beings. Human brain-AI symbiosis would be the answer to the existential risk to humanity posed by the superintelligent AI.       

An artificial intelligence (AI) system is a language model (LM) that makes a probabilistic prediction of the next word in a natural language given preceding one(s). The model is pre-trained with the data so that it predicts what comes next in the sentences when prompted. In doing so, the model mimics the function of natural intelligence.   

Older forms of AI modeled reasoning. They were based on the idea that the essence of human intelligence is reasoning or logic. According to this symbolic approach, the meaning of a word is how it relates to the other words. Understanding a sentence meant translating the sentence into some internal symbolic language. Then, one applies rules to the symbolic expressions to derive new expressions. The early intelligent systems based on this idea were not very effective, and no significant headway could be made in the discipline although AI had its beginnings as early as the 1950s.  

Massive progress has been made in AI in recent years. New forms of AI have come into existence which are highly efficient. Many factors have worked together to make that happen, one of which is an emphasis on a biological or psychological approach to human intelligence and how human brain functions. According to biological approach, the meaning of a word is a set of properties or features and understanding means converting each word symbol into a bunch of features. The new AI forms unify the two approaches. It converts each word into a big set of features. Interactions between the features of different words enable prediction of features of the next word which in turn enables prediction of the next word given its features.  

New AI forms model human intuition (instead of reasoning). They are based on neural network and process data in a way similar to the human brains. A large-scale neural network language model performs a variety of natural language processing tasks in an efficient way. Important current Large Language Models (LLMs) like xAI’s Grok, Google’s Gemini, Anthropic’s Claude, OpenAI’s ChatGPT, High-Flyer’s DeepSeek and others have vast computing powers. They are very well trained and highly efficient. Their unmatched computing power has impacted many areas. There is report of Anthropic’s Claude being used for analysis, pattern identification, planning, simulation, war gaming in currently ongoing war in the Middle East region.   

Brain-Computer Interfaces (BCI) technology is one such area that has immensely benefited from recent developments in AI. The technology is not new but the vast computing power of latest LLMs has eased decoding and processing of neural signals. As a result, many BCI devices have now reached clinical trial stage.  

Neuralink, one of the important players in the field, is developing a brain implant, a brain-computer interface (BCI) namely “Telepathy” that will enhance autonomy and independence of people with debilitating conditions such as spinal injury, stroke, ALS, etc. It will enable such people to directly control computers, phones, and assistive devices like robotic limbs using their thoughts alone (telepathy, in behavioural science, refers to the parapsychological phenomenon involving direct communication of thought from one person’s mind to another person’s mind without using the usual sensory channel and any known signals). This BCI device is currently undergoing three early feasibility trials. While PRIME Study involving 15 participants is testing neuronal control of external devices, the CONVOY study on three participants is investigating control of assistive devices and VOICE Study on 6 participants is exploring restoration of phonation reminding of how Stephen Hawking was shown to communicate in the television sitcom “Big Bang Theory”. Neuralink’s other brain implant “Blindsight”, a vision-restoring implant is in the pipeline for clinical trial awaiting regulatory approval. 

The BCI medical devices being developed by Neuralink replace the damaged biological neuronal interfaces and restore natural and intuitive interactions with the digital and physical worlds for those with unmet medical needs. The Telepathy device picks up the command signal from the brain and delivers to the external effectors like computer, phone or assistive device for execution of the task. The Blindsight device, on the other hand, will process the sensory signals collected from the external environment for visual perception by the brain. In this case, signals from the external environment will be converted to neural signals with the help of AI and fed into the visual cortex for perception bypassing the damaged sensory interface. Decoding and processing of signals have become possible courtesy of modern LLMs. Success is also due to 1024-channel implant which has vastly improved data transfer rate from brain to the computer. Though still in clinical trial stage, these BCI implants will vastly improve the quality of life of affected people when commercialised in near future. However, there is more to the story of advances in BCI technology.    

In the above-mentioned clinical trials, AI is being used for decoding and processing of neural signals collected by the implants in the brains of individuals with the unmet needs where the brain bypasses the damaged biological interfaces and communicates with the external computer directly. Can an otherwise healthy individual use the vast computing power of AI platforms in a similar way to enhance efficiency and performance to become superhuman? 

Here is an excerpt from what physicist Michio Kaku said about AI while discussing technologies of the future in 2018:  “…I think that the tipping point as to when robots become dangerous is when they have self-awareness, perhaps by the end of the century. Right now, our most advanced robots have the intelligence of a cockroach — a retarded lobotomized cockroach. But eventually our robots will become as smart as a mouse, then as smart as a rat, then a rabbit, then a dog and a cat, and by the end of this century, perhaps as smart as a monkey. At that point, they are potentially dangerous. Monkeys know they are monkeys. Monkeys know they are not human. Now, dogs are confused. Dogs don’t know that we are not dogs. Dogs think that we are dogs and therefore they obey us — we’re the top dog, they are the underdog. So I think, at that point, a hundred years from now, at the end of the century, we should put a chip in their brain to shut them off if they have murderous thoughts. That is a fail-safe mechanism, but that’s only temporary because then what happens when robots become so smart they remove the fail-safe system? That is also possible in the next century, the 22nd century. At that point, I think we should merge with them. I don’t think this will happen in this century, but I think in the next century we should merge with our creation. Why not become Homo superior? Why not use exoskeletons which are now being created to become Hercules? That is the power of a god. So, in other words, an option instead of fighting off the robots, in the next century, is to merge with them to become superhuman…” — Michio Kaku (2018)Technologies of the Future.

Since Michio Kaku made the above observation in 2018 that in the future, “man will merge with the robots to become superhuman”, Brain-Computer Interfaces (BCI) technology seems to be progressing towards that prediction courtesy of advances in compute capacity of artificial intelligence (AI) systems. 

The primitive limbic system of our brain (the emotional brain) is the source of purpose for most of us most of the time. Our cerebral cortex (the thinking and planning brain) uses a massive amount of compute as a secondary layer to serve the limbic system. In doing so, the cortex is augmented by a tertiary computing layer comprising phones, laptops, iPad, and applications including AI platforms to enhance performance. The brain in this case communicates with the tertiary compute layer through our biological interfaces either by typing or speaking where the rate of data transfer from the cortex to the tertiary compute layer is abysmally low hence a bottleneck. Can human brains communicate with the AI platforms at high speeds characteristic of superintelligent AI computing systems?   

A high-speed connection that allows for a high-fidelity data stream directly into the cerebral cortex from the AI (and vice versa from cortex to AI) would help effectively integrate AI in our tertiary compute layer. This is exactly what is happening in the above-mentioned clinical trials – Neuralink’s Telepathy implants establish a high-speed connection between brain (of the people with unmet medical needs) and the computer bypassing the damaged biological interfaces thus integrating AI in their tertiary compute layer. As a result, the trial participants are able to use phones and computers to browse the internet, to send messages and to compose emails, play video games, and use robotic limbs for chores involving manual dexterity through thoughts alone. The new ability is vastly improving the quality of life of the participants. From a technological point of view, integrating AI in our tertiary compute layer to augment function through a high bandwidth connection between the brain and the computer (replacing our slow biological interfaces) is a milestone. 

Of course, there is a robust case for pursuing the technology for meeting the medical needs, but what about integrating AI in our tertiary compute layer for augmenting functions among otherwise healthy people? The technology is not very far away; it is already undergoing human trial albeit on people with unmet medical needs. But will it stop there?   

Ironically, AI is already in our tertiary compute layer along with all other computing stuff and is augmenting functions to the extent our slow biological interfaces could allow. We transmit data at roughly 10 to 100 bits per second (bps), average over 24 hours is about 1 bit per second (bps). So, we interact with the AI platforms through our extremely slow biological interfaces which are bottlenecks in the brain’s communication with the superintelligent AI. There is, thus, a gross mismatch – we can transmit roughly 10 to 100 bits per second while current AIs can process and output trillions of bits per second. This means our ability to communicate our intent to AI, and AI’s ability to feed complex insights back into our consciousness, is throttled by our biology. Consequently, the two (viz. Brain and the AI) remain outside of each other. Clearly, humans run the risk of being obsolete in the face of superintelligent AIs. There is an existential risk to humanity. Can AI be stopped in view of risks? Looks unlikely because it has a strong economic rationale for corporates in terms of operational cost reduction and profit enhancement. More importantly, AI has already found significant applications in national security, defence, and war. Outcome of any future war would be critically dependent on augmentation of defence capabilities through AI; hence state agencies would strive for AI capacity building. This makes AI indispensable to countries for national defence.  

Current trends in technological advances indicate that soon it may become possible to establish high-speed connection between the brain and AI platforms bypassing abysmally slow biological interfaces to effectively integrate AI into our tertiary compute level. The high-bandwidth neural links would serve as a bridge, effectively merging the brain with AI. Humans would become cyborgs (cybernetic organisms). The merger would enable the two to benefit from each other. The brain would acquire superhuman computing power of AI thereby mitigating the risk of humans being obsolete in the face of superintelligent digital beings. Human brain-AI symbiosis would enable humans to control AI, thus be the answer to the existential risk to humanity posed by the superintelligent AI.    

*** 

Sources:  

1. StarTalk (28 February 2026). Is AI Hiding Its Full Power? With Geoffrey Hinton. Available at https://www.youtube.com/watch?v=l6ZcFa8pybE 
2. Canada Info ((27 February 2026)). WE’RE TOAST: Godfather of AI Geoffrey Hinton warns Canada’s Senate of EXISTENTIAL threat to humanity. Available at https://www.youtube.com/watch?v=7fImPlfdRS0 
3. Neuralink. Updates – Two Years of Telepathy. Posted 28 January 2026. Available at https://neuralink.com/updates/two-years-of-telepathy/ 
4. PRIME: An Early Feasibility Study of a Precise Robotically Implanted Brain-Computer Interface for the Control of External Devices. Available at  https://clinicaltrials.gov/study/NCT06429735
5. CONVOY: An Early Feasibility Study of Neural Control of Assistive Devices Via Brain-Computer Interface Technology. Available at https://clinicaltrials.gov/study/NCT06710626  
6. VOICE: An Early Feasibility Study of a Precise Robotically Implanted Brain-Computer Interface for Communication Restoration. Available at https://clinicaltrials.gov/study/NCT07224256 
7. Lex Fridman (2 August 2024). Elon Musk: Neuralink and the Future of Humanity. Lex Fridman Podcast #438. Available at https://www.youtube.com/watch?v=Kbk9BiPhm7o 
8. Kumar, R., Waisberg, E., Ong, J., & Lee, A. G. (2025). The potential power of Neuralink – how brain-machine interfaces can revolutionize medicine. Expert Review of Medical Devices, 22(6), 521–524. https://doi.org/10.1080/17434440.2025.2498457  
9. Bandre, P., et al 2025. “Neuralink: Revolutionizing Brain–Computer Interfaces for Healthcare and Human–AI Integration,” 2025 2nd International Conference on Electronic Circuits and Signaling Technologies (ICECST), Petaling Jaya, Malaysia, 2025, pp. 1122-1126, DOI: https://doi.org/10.1109/ICECST66106.2025.11307276 
10. UC Davis Health. New brain-computer interface allows man with ALS to ‘speak’ again. 14 August 2024. Available at https://health.ucdavis.edu/news/headlines/new-brain-computer-interface-allows-man-with-als-to-speak-again/2024/08 
11. Vansteensel MJ, et al 2016. Fully Implanted Brain–Computer Interface in a Locked-In Patient with ALS. N Engl J Med. 2016 Nov 12;375(21):2060–2066. DOI: https://doi.org/10.1056/NEJMoa1608085 
12. Zhang X., et al 2020. The combination of brain-computer interfaces and artificial intelligence: applications and challenges https://doi.org/10.21037/atm.2019.11.109 

*** 

Related articles:  

PRIME Study (Neuralink Clinical Trial): Second Participant receives Implant (8 August 2024)  

Neuralink: A Next Gen Neural Interface That Could Change Human Lives (29 August 2020)  

BrainNet: The First Case of Direct ‘Brain-To-Brain’ Communication (5 July 2019) 

***