Neuralink and the Future of the Human Nervous System
AlphaGo, a project created by an artificial intelligence company called Deepmind, is an AI designed to master the art of Go. Go, an ancient abstract strategy board game invented in China, is avowed for its complexity and difficulty. The intricacies of the game made it even more astonishing that AlphaGo was able to defeat the Go World champion, Lee Sedol 4-1 in March 2016. What’s even more extraordinary, is that the next iteration of the project, AlphaGo Zero, was able to defeat the original AlphaGo at the game 100-0 in 40 days without any form of human interaction. As we advance technologically, there may come a point where we may represent the minority of intelligence – putting us, as a species, at risk.
The AI present today is considered narrow AI, also known as weak AI. Narrow AI is named such because its sole purpose is to manage one task. AlphaGo, image/speech recognition algorithms, Daily Mixes on Spotify, and recommendations in Netflix and Amazon are all forms of narrow AI. These examples function by the AI’s ability to train itself at the job in hand until it becomes intelligent in that task. This process, which is alike to how infants learn, is termed machine learning. When we were infants, we bettered our learning of the environment by digesting more data and information from our observations and interactions. As problems become more complex, it becomes more difficult for humans to solve. This is where machine learning shines. Not only can it analyze data, but the computer can also learn from it and adapt its own view efficiently. This is what allowed AlphaGo to surpass thousands of years of human strategy in only 40 days.
Although current artificial intelligence systems are not an immediate threat as of yet, the concern raises when Artificial General Intelligence (AGI) is introduced. As opposed to narrow AI, AGI has more than one purpose. Although humans may not be able to process information at the speed of light, we have the unique ability to process random, unrelated thoughts and observations and still come to a solution to a complex problem. Humans don’t have to brute force their way through every option to learn how to solve a problem as much as a machine does. This type of thinking allowed us to create societies, play games, and laugh. These human traits and ideals are difficult to teach a computer. Also, because of machine learning, AI’s intelligence would increase exponentially – much faster than the growth of human intelligence. This would suggest that in a short amount of time, the AGI would become smarter than all of humanity. As some may argue, this could propose a problem for our species.
As opposed to discontinuing our technological advancements with artificial intelligence, some, most notably Elon Musk, suggest that the better, more pragmatic solution would be to merge with them and become “cyborgs”. This idea is not as far fetched as it may initially seem. According to Musk, to some degree, we are already cyborgs. A cyborg is defined as a person whose physical abilities are extended beyond normal human limitations by mechanical elements. Although it may not be obvious, our smartphones help us fit into this definition. Smartphones have become an extension of ourselves. They help us remember everything from our notes, photos, and videos. They also enable us to answer any question at a moment’s notice with the aid of the internet. The only issue is that the inputs are too slow.
Musk provides a solution to this with his company, Neuralink, by proposing a high bandwidth link between our brains and the internet using what he calls a Neural Lace. The human nervous system is categorized into two major components: the limbic system and the cortex. The limbic system is responsible for emotional intelligence and survival instincts while the cortex is responsible for problem solving and critical thinking. Neuralink’s vision is to incorporate a third component that allows the AI to increase our capabilities to multiple orders of magnitude by giving us perfect memory, access to all information available on the internet, and more.
The process to carry out this goal is still being determined. There have been multiple brain machine interfaces (BMIs) created already. A method that was done was implanting an electrode lined neural mesh into the brain. This may sound frightening, but the process itself is not very invasive. The process (already tested on mice) involves a needle where the mesh would unfold itself from the syringe once injected into the skull. Another method done was fitting a million electrodes into a device that is about the size of two nickels stacked and implanting it into the skull. Another popular method known as transcranial magnetic stimulation (TMS), which involves having a magnetic coil outside the head that allows electric impulses inside of the brain. Although these methods are already advanced, Neuralink plans to improve upon these methods using what they call Neural Dust. These “dust” particles are actually 100-micrometer sensors that are about the width of a strand of hair. These particles will then be sprinkled through the cortex where a three mm-size device will communicate with these sensors using ultrasound technology. The process includes using microchip technology as well as RFID.
The plans at Neuralink are not yet finalized. There has been discussion on using optosomes (viruses attached to brain cells that are stimulated by light) and carbon nanotubes (bundled and sent to the brain via the bloodstream) in order to incorporate the AI into our brains. Nevertheless, with artificial intelligence advancing exponentially, its threat may be imminent. Instead of pulling the plug, the best step for humanity might just be to incorporate artificial intelligence into our biology to allow us to keep our control and accomplish feats at a rate that was previously unimaginable.
References:
http://sciencenetlinks.com/science-news/science-updates/checkers-solved/
https://deepmind.com/blog/alphago-zero-learning-scratch/
https://medium.com/point-nine-news/what-does-alphago-vs-8dadec65aaf
https://skymind.ai/wiki/strong-ai-general-ai
https://skymind.ai/wiki/neural-network
https://waitbutwhy.com/2017/04/neuralink.html
https://www.popsci.com/science/article/2010-12/fyi-how-many-different-ways-can-chess-game-unfold
https://www.youtube.com/watch?v=Pls_q2aQzHg