A commentary on the plan to control computers with minds

This blog is one in our series surrounding some of the more general AI topics. These are not directly related to Conversational AI, but make interesting reading.

What if you could control a computer or a machine with your mind? This question has featured in many science-fiction novels throughout the decades, from the mind-controlled industrial equipment and spaceships of Samuel Delany’s Nova to the fully immersive cyberspace of William Gibson’s Neuromancer. (Note of interest: one of action.ai’s server clusters is named neuromancer.)

However, the most prophetic writer in this regard is lesser known. In Pat Cadigan’s novel Synners, an eccentric billionaire commissions a research project to breach the computer-brain barrier by drilling holes and inserting wires into the human skull. Such a plotline holds eerie parallels with one of Elon Musk’s latest ventures, Neuralink.

Front cover of Cadigan’s Synners.

What is Neuralink?

Founded by Musk in 2016, Neuralink is a neurotechnology company that aims to develop a brain-machine interface. This device, called the Link, will be a small implant inserted into the brain via a hole drilled into the skull. Incredibly small, flexible threads that each contain numerous electrodes will be attached to neurons in the brain and monitor their signals. The signals will then be fed back into a computer chip which will allow the user to control external machines using their thoughts.

Prototype of the Link.

Because these threads are so delicate they cannot be inserted by a human surgeon. To overcome this Neuralink are developing a surgical robot that will be able to reliably insert the implant and its corresponding threads. The implant itself holds a battery that will be charged wirelessly with a conductive charger. The company wants to make the implant as discreet as possible, so it will not be externally visible on the human body. Currently, Neuralink aims to develop a mobile app that will allow the user to control their phone’s touch screen and keyboard using the Link.

How it works

The Link implant aims to record and stimulate brain activity in order to enable the remote manipulation of electronic devices. If, for example, the human brain produces the same repeatable pattern when thinking about moving a limb, or using that limb to move a computer mouse, then theoretically an implant could detect and interpret these brain patterns before wirelessly signalling its intent to a computer. This computer would in turn perform the action that the user is thinking about. In the words of Neuralink themselves:

We are designing the Link to connect to thousands of neurons in the brain, so that it may one day be able to record the activity of these neurons, process these signals in real-time, and translate intended movements directly into the control of an external device.

In multiple scientific and technological fields, including materials, neuroscience, and artificial intelligence, Neuralink is pushing against what is currently feasible and theoretically possible. Numerous barriers still remain before the Link can reach the market:

The electrode threads need to be durable and resistant to corrosion. This necessitates the development of new corrosion-resistant adhesive materials.
The chip must be powerful enough to detect human brain patterns, but it also needs to be power efficient so it can operate from a small battery carried inside the human brain.
As mentioned, a surgical robot needs to be developed to actually implant the threads.

Machine learning algorithms are also required to decode the information received from neurons in the brain. Highly advanced algorithms will need to be developed to map brain activity at this scale and with the necessary level of detail.

Use cases

Like many recent advances in artificial intelligence, robotics, and medical science, the Link has a wide range of potentially groundbreaking use cases if used responsibly and within a robust ethical and regulatory framework.

For instance, the new tech could help humans control prosthetic limbs with a high degree of precision and accuracy. This would have clear benefits for those who are paralysed or who suffer from motor function disabilities. These groups could use the Link to operate computers and machines more effectively.

More broadly, any individual could use the Link to control machines and computers more efficiently using just their thoughts. This would likely have wide-ranging applications across manufacturing, transport, logistics, surgery, and even virtual entertainment.

Inside the Link.

Neuralink also hopes to treat neurological disorders using their implants. While the specifics of this are vague, the device might effectively function like a pacemaker for the brain and regulate its functions. While this holds enormous therapeutic potential, this application especially requires strong regulation against misuse.

Ultimately, Neuralink aims to breach the human-machine barrier and is a step towards a world in which human consciousness moves seamlessly across organic matter and the silicon chip. If Neuralink’s ambitions are realised the implications for human society and identity are profound.

What are the risks?

The technology pushes at the limits of neuroscience and our understanding of the human brain. Placing any kind of implant into the brain is risky, and it is difficult to predict the long-term effects of something like the Link. Additionally, putting any internet-capable device into direct contact with the human brain carries enormous security risks that need to be taken seriously.

Even with modern neuroscience, the brain’s functions are still not completely understood. It’s not even clear on a theoretical level if a brain-machine interface would even work. The major challenge comes from the AI modelling necessary to interpret the function of each individual neuron. Is it really possible to map out the functions of the brain’s entire neural network and build a model sophisticated enough to accurately predict its behaviour? And would such patterns be consistent and recognisable from one individual to another? Reading surface-level functions from brain activity is already a monumental challenge and accessing deeper thoughts, memories, and emotions would require a paradigm shift in neuroscience.

Furthermore, the project has already attracted considerable controversy, most notably allegations of animal cruelty. This year, the Physicians Committee for Responsible Medicine accused Neuralink of subjecting several monkeys to psychological distress and extreme suffering. Neuralink itself has admitted that monkeys have died as a result of the project, but they deny that any abuse has occurred.

If Neuralink ever reaches clinical trials or develops a commercial offering, it’s almost certain that controversy won’t be far behind. And this could put significant pressure on the scientific community and lawmakers to regulate the new technology.

Beyond the hype

Regardless of the hype surrounding it, there are still reasons to be optimistic about Neuralink. The company’s current prototype has been successfully inserted into, and removed from, the brain of a pig and contains 1,024 flexible electrodes or threads. For context, the closest current equivalent to the Link used in medical science, the Utah Array, contains just 64 electrodes and cannot be inserted or removed without considerable risk of tissue damage.

It’s too early to tell whether Neuralink and Musk will achieve their end goals. But the company is pushing in enough directions that advances in materials, neurosurgery, robotics, and artificial intelligence are likely to follow. While the science-fiction visions behind the project are still a long way from reality, hopefully the project will take us one step closer to learning how the enigmatic human brain really works.