Elon Musk wants to hack the brain.
The CEO of SpaceX and Tesla has launched a medical research company called Neuralink in California, the Wall Street Journal reported. The goal is to create brain-computer interfaces that would allow humans to connect directly with what he’s called “the digital version of yourself” — electronic devices.
Musk has been teasing the possibility of a venture into computer-brain interfaces for months. He spoke at Recode’s Code Conference in 2016 about the need to speed up human “output.” Essentially, he said, humans are incredibly rapid at taking in information, but slow at outputting information to their digital devices. At the World Government Summit in Dubai, the United Arab Emirates, in January, Musk called artificial intelligence “dangerous” because it could render humans obsolete.
“This is going to be a massive social challenge,” he said at the conference.
Musk advocated a universal basic income, or a basic payment to unemployed people around the world, to confront these challenges. But he also floated the idea of a “merger with biological intelligence and machine intelligence.”
“To some degree, we are already a cyborg,” Musk said at the Dubai conference. “You think about the digital tools you have, your phone, your computer, the applications that you have. … You already have a digital tertiary layer.”
Currently, people interact with their devices by thumb-typing on their phones, Musk said. A “high-bandwidth interface to the brain” would help achieve a symbiosis between human and machine intelligence and could make humans more useful in an AI-driven world, he said.
In science fiction, this idea is sometimes called “neural lace” for the netting of electronic implants that would presumably be required for such an interface.
Brain-human interfaces must overcome steep challenges, though. So far, there have been a few successful brain-implant devices in humans, all designed to treat serious neurodegenerative conditions or neurological injuries. Deep-brain stimulation, electrical pulses delivered into the brain, is sometimes used to slow the symptoms of Parkinson’s disease when medications fail to work, for example.
Several patients with spinal cord injuries have been fitted with implants giving them some control over robotic limbs, or even their own limbs. This represents a leap from animal tests to experimental use on humans in less than a decade, according to a paper in the journal Frontiers in Systems Neuroscience. Nevertheless, wrote Gytis Baranauskas, a neurophysiologist at the Lithuanian University of Health Sciences, the rate by which electronic systems transfer impulses from the brain to the limb or prosthetics (or vice versa) lags far beyond natural nerve impulses, especially for complex movements. It’s not the electronic technology that limits this transfer of information, Baranauskas wrote in the Frontiers paper, but a lack of understanding of what neuron activity in the brain really means.
In other words, there is a lot more neuroscience to do before anything close to neural lace becomes a reality. There are also risks to consider: It’s one thing to chance a 1 percent to 3 percent likelihood of a brain bleed, stroke or infection to implant an electrode in an effort to slow a fatal disease like Parkinson’s. It’s another thing altogether to consider surgery that would probably be far more invasive so that people could better control their computers.