ScienceSocietySpiegeloog 405: Together

Merging Minds and Machines

By June 19, 2020 No Comments

What’s the fundamental difference between your brain, and the computer you are looking at, at this very moment? Right now, the gap between them seems huge, but this is starting to change. Elon Musk’s company Neuralink attempts to bring brains and computers closer together – with potentially mind-blowing consequences.

What’s the fundamental difference between your brain, and the computer you are looking at, at this very moment? Right now, the gap between them seems huge, but this is starting to change. Elon Musk’s company Neuralink attempts to bring brains and computers closer together – with potentially mind-blowing consequences.

Image: Gerd Altmann 

Can computers think, and if so, how is it different from human thinking? Humans think about lots of things. They think about which career to pursue, they try to remember the groceries they want to buy from the supermarket, or they might think about which move to make in a game of chess. While a computer cannot make important life decisions for you, it very well can help you with the latter two tasks. As we probably all know, a computer can remember most information better than any living human, and with the right programs installed, it will always beat the best chess player. In more nerdy terms, the computer has some form of what’s called ‘narrow’ AI (Artificial Intelligence), which means that there are highly specialized tasks (e.g. chess) it is good at, but if you ask it to do anything else, it will fail (Bostrom, 2014). So, does the computer think when it is calculating which chess move to play next? As the Dutch computer scientist Edsger Dijkstra has noted, computers are like submarines in this regard. Do submarines swim? Well, they do something similar, and if you want to call this swimming – fine. In the same way, computers do something that is somewhat related to what humans do, but whether this already counts as thinking is a matter of definition (Beniaminy, n..d.). The critical point here, though, is that a computer – no matter whether it can ‘think’ or not – could currently be viewed as an extension of the human mind, rather than having a mind on its own. If you often forget to go grocery shopping, a computer (e.g. your smartphone) can remind you to do so; and if you are a bad chess player, the computer can play the game for you. A computer is therefore a tool that can do part of the ‘thinking’ for you, without being able to replace you.

“the best prosthetics today are already sophisticated enough to be controlled by thought alone”

Why is this relevant? After conceptualising a computer as something that can extend the human mind, it becomes clear how inefficient normal interactions between minds and computers are: Considering the speed of the electronic and biochemical circuits operating in computers and brains, using something like your fingers to interact with a computer is incredibly slow – and this is exactly where Neuralink comes into play. Musk’s company aims to develop high bandwidth brain-machine interfaces that would enable humans to connect their minds to the computer directly (Musk, 2019).

Within brain-machine interfaces, there are two important sub-goals that researchers are trying to accomplish: extracting the right information from the brain to the computer and transmitting the right information from the computer into the brain. For example, the best prosthetics today are already sophisticated enough to be controlled by thought alone (Rosen & Ferguson, 2020). This is usually done by placing a tiny electrode into the motor cortex of the person, which is able to record the specific neuronal activity that controls the limb. Hence, the right information in the brain can be extracted and sent to the artificial limb, which can then be moved. However, there are hardly any artificial limbs that are also able to send the right information (i.e. sensory information) into the brain, and so current artificial limbs will never feel like real parts of one’s body. Another, albeit different, example is cochlear implants for people with hearing impairments. In contrast to prosthetic limbs, transmitting the correct information into the brain is already possible in this domain: Cochlear implants are able to transform acoustic signals into neuronal information, which the brain then uses to create the sound the person with the implant ‘hears’. If these two problems – extracting information from the brain and transmitting information into the brain – can be sufficiently solved (i.e. with higher resolution, without requiring dangerous surgeries, etc.), then Neuralink’s goal of merging brains and machines is already incredibly close.

“I could simply send you the mental image, and you would know exactly what I am thinking of.”

It’s hard to imagine the full scope of the consequences of Neuralink’s success. Prosthetic limbs would truly feel like a part of you. People who have lost their hearing or eyesight could regain the senses they once possessed. Neither a keyboard nor voice recognition will be necessary to make a google search anymore. Any kind of knowledge could be learned without ever reading a book, and at a much faster rate than ever before. Furthermore, artificial intelligence, as it is improving throughout the next decades, would complement human intelligence through brain-machine interfaces, and help humans expand their cognitive potential in unprecedented ways. There are so many things we are simply not smart enough to understand, but it looks like with Neuralink our minds could truly reach for the stars.

There is more to Neuralink. Just like the Internet enables you to communicate with almost anyone on the planet at any time, Neuralink would make it possible to transmit this information directly from person to person. Let’s say I am picturing something that doesn’t exist, like a fish-squirrel hybrid, in my head. In today’s world, if I want you to know what it looks like I would need to describe in detail what I am thinking of, and even then the picture you have in your mind will be very different from mine. In Neuralink’s world, there is no need for describing my imagination orally. I could simply send you the mental image, and you would know exactly what I am thinking of. Similarly, let’s consider what it feels like split seconds before you are having a dangerous traffic collision. It’s such an intense feeling that someone who hasn’t experienced it could hardly understand what it is like. Yet, with Neuralink, you could transmit the same feeling to another person, and this person would understand exactly what you had gone through. Finally, consider what would happen if two children are raised while an efficient brain-machine interface is keeping them constantly connected. If you always feel what your sibling feels, to what extent are you still a different person from him or her?

“Some of the thoughts presented here might seem indeed a bit like science fiction”

Needless to say, all of this is in the distant future. Is it even possible? Some of the thoughts presented here might seem indeed a bit like science fiction, and yet there is in theory no barrier that would prevent Neuralink from achieving its goals. And although I personally believe that Neuralink’s tremendous progress (i.e. the fact that Neuralink’s products are already far better than anything that currently exists in the market; Alexander, 2020) shows that it can indeed be done, you could argue that it still sounds too good to be true. Ultimately, no one knows what the future will bring. In this way, the fact that Musk pumps millions of dollars into his idea could mean he’s crazy – or that an incredible future is awaiting us.

References

– Alexander, D. (2020, March 16). Downloading the Human Brain to a Computer: Elon Musk’s Neuralink. Interesting Engineering. Retrieved from: https://interestingengineering.com/downloading-the-human-brain-to-a-computer-elon-musks-neuralink.
– Beniaminy, I. (n.d.). Do we Think they Think? The Future of Things. Retrieved from: https://thefutureofthings.com/3686-do-we-think-they-think/.
– Bostrom, N. (2014). Superintelligence: Paths, Dangers, Strategies. Oxford University Press.
– Musk, E. (2019). An integrated brain-machine interface platform with thousands of channels. Journal of Medical Internet Research, 21(10).
– Rosen, J. & Ferguson, P. W. (2020). Wearable Robotics. London: Academic Press.

Can computers think, and if so, how is it different from human thinking? Humans think about lots of things. They think about which career to pursue, they try to remember the groceries they want to buy from the supermarket, or they might think about which move to make in a game of chess. While a computer cannot make important life decisions for you, it very well can help you with the latter two tasks. As we probably all know, a computer can remember most information better than any living human, and with the right programs installed, it will always beat the best chess player. In more nerdy terms, the computer has some form of what’s called ‘narrow’ AI (Artificial Intelligence), which means that there are highly specialized tasks (e.g. chess) it is good at, but if you ask it to do anything else, it will fail (Bostrom, 2014). So, does the computer think when it is calculating which chess move to play next? As the Dutch computer scientist Edsger Dijkstra has noted, computers are like submarines in this regard. Do submarines swim? Well, they do something similar, and if you want to call this swimming – fine. In the same way, computers do something that is somewhat related to what humans do, but whether this already counts as thinking is a matter of definition (Beniaminy, n..d.). The critical point here, though, is that a computer – no matter whether it can ‘think’ or not – could currently be viewed as an extension of the human mind, rather than having a mind on its own. If you often forget to go grocery shopping, a computer (e.g. your smartphone) can remind you to do so; and if you are a bad chess player, the computer can play the game for you. A computer is therefore a tool that can do part of the ‘thinking’ for you, without being able to replace you.

“the best prosthetics today are already sophisticated enough to be controlled by thought alone”

Why is this relevant? After conceptualising a computer as something that can extend the human mind, it becomes clear how inefficient normal interactions between minds and computers are: Considering the speed of the electronic and biochemical circuits operating in computers and brains, using something like your fingers to interact with a computer is incredibly slow – and this is exactly where Neuralink comes into play. Musk’s company aims to develop high bandwidth brain-machine interfaces that would enable humans to connect their minds to the computer directly (Musk, 2019).

Within brain-machine interfaces, there are two important sub-goals that researchers are trying to accomplish: extracting the right information from the brain to the computer and transmitting the right information from the computer into the brain. For example, the best prosthetics today are already sophisticated enough to be controlled by thought alone (Rosen & Ferguson, 2020). This is usually done by placing a tiny electrode into the motor cortex of the person, which is able to record the specific neuronal activity that controls the limb. Hence, the right information in the brain can be extracted and sent to the artificial limb, which can then be moved. However, there are hardly any artificial limbs that are also able to send the right information (i.e. sensory information) into the brain, and so current artificial limbs will never feel like real parts of one’s body. Another, albeit different, example is cochlear implants for people with hearing impairments. In contrast to prosthetic limbs, transmitting the correct information into the brain is already possible in this domain: Cochlear implants are able to transform acoustic signals into neuronal information, which the brain then uses to create the sound the person with the implant ‘hears’. If these two problems – extracting information from the brain and transmitting information into the brain – can be sufficiently solved (i.e. with higher resolution, without requiring dangerous surgeries, etc.), then Neuralink’s goal of merging brains and machines is already incredibly close.

“I could simply send you the mental image, and you would know exactly what I am thinking of.”

It’s hard to imagine the full scope of the consequences of Neuralink’s success. Prosthetic limbs would truly feel like a part of you. People who have lost their hearing or eyesight could regain the senses they once possessed. Neither a keyboard nor voice recognition will be necessary to make a google search anymore. Any kind of knowledge could be learned without ever reading a book, and at a much faster rate than ever before. Furthermore, artificial intelligence, as it is improving throughout the next decades, would complement human intelligence through brain-machine interfaces, and help humans expand their cognitive potential in unprecedented ways. There are so many things we are simply not smart enough to understand, but it looks like with Neuralink our minds could truly reach for the stars.

There is more to Neuralink. Just like the Internet enables you to communicate with almost anyone on the planet at any time, Neuralink would make it possible to transmit this information directly from person to person. Let’s say I am picturing something that doesn’t exist, like a fish-squirrel hybrid, in my head. In today’s world, if I want you to know what it looks like I would need to describe in detail what I am thinking of, and even then the picture you have in your mind will be very different from mine. In Neuralink’s world, there is no need for describing my imagination orally. I could simply send you the mental image, and you would know exactly what I am thinking of. Similarly, let’s consider what it feels like split seconds before you are having a dangerous traffic collision. It’s such an intense feeling that someone who hasn’t experienced it could hardly understand what it is like. Yet, with Neuralink, you could transmit the same feeling to another person, and this person would understand exactly what you had gone through. Finally, consider what would happen if two children are raised while an efficient brain-machine interface is keeping them constantly connected. If you always feel what your sibling feels, to what extent are you still a different person from him or her?

“Some of the thoughts presented here might seem indeed a bit like science fiction”

Needless to say, all of this is in the distant future. Is it even possible? Some of the thoughts presented here might seem indeed a bit like science fiction, and yet there is in theory no barrier that would prevent Neuralink from achieving its goals. And although I personally believe that Neuralink’s tremendous progress (i.e. the fact that Neuralink’s products are already far better than anything that currently exists in the market; Alexander, 2020) shows that it can indeed be done, you could argue that it still sounds too good to be true. Ultimately, no one knows what the future will bring. In this way, the fact that Musk pumps millions of dollars into his idea could mean he’s crazy – or that an incredible future is awaiting us.

References

– Alexander, D. (2020, March 16). Downloading the Human Brain to a Computer: Elon Musk’s Neuralink. Interesting Engineering. Retrieved from: https://interestingengineering.com/downloading-the-human-brain-to-a-computer-elon-musks-neuralink.
– Beniaminy, I. (n.d.). Do we Think they Think? The Future of Things. Retrieved from: https://thefutureofthings.com/3686-do-we-think-they-think/.
– Bostrom, N. (2014). Superintelligence: Paths, Dangers, Strategies. Oxford University Press.
– Musk, E. (2019). An integrated brain-machine interface platform with thousands of channels. Journal of Medical Internet Research, 21(10).
– Rosen, J. & Ferguson, P. W. (2020). Wearable Robotics. London: Academic Press.
Valentin Weber

Author Valentin Weber

Valentin Weber is a third year PPLE student, interested in the intersection of philosophy, neuroscience, and psychology. He also enjoys playing chess, solving riddles, and heated discussions.

More posts by Valentin Weber