If anyone understands the unique level of risk associated with getting a neural implant, it’s Dr. Phil Kennedy. In 2014, he was forced to fly to Belize to have a surgeon there perform an experimental brain surgery the FDA had deemed too dangerous for approval. Kennedy’s goal was to wire his brain with intracranial electrodes of his own design, with a plan to awaken with a simple, working interface between his brain and a computer, and a lifetime of all-new possibilities.
Instead, he awoke to find himself far more limited than he had been before. His movements were imprecise, and his mouth simply would not form the words he wanted. Dismayed, Kennedy stumbled through the first few days heavily disabled, plagued by headaches and other complications. It seemed obvious that he had made a huge mistake but, having put the devices inside his skull, what was there to be done about it now?
With months of effort, Kennedy was eventually able to regain his full powers of movement and speech, and in a recent phone-call with Inverse he made it clear that the experience hasn’t dulled his enthusiasm for the technology in the slightest. When asked how, given his ordeal, he feels about the prospect of having millions of people receive neural implants in the near future, he replied that a neurally wired world can’t come quickly enough.
“Even for people who are able-bodied, controlling a computer and accessing the cloud directly … I think it’s a great idea,” he said, with totally unironic enthusiasm. “I’m totally for it. It’s my ultimate goal!”
It’s the ultimate goal of a growing number of others in the tech world, too. There’s exciting research coming out of Facebook in particular, but also smaller projects from the likes of Mattel and even Kickstarter entrepreneurs. These all opt for the much easier option of reading the brain without surgery, however, and as a result their potential for reading and manipulating the brain is much lower than techniques that put devices directly onto the surface of the brain. The assumption with many these days, and certainly with Elon Musk, is that these non-invasive devices will be inevitably followed by a shift toward Kennedy’s preferred intracranial future.
Yet, as Kennedy knows better than most, all sorts of unique new difficulties arise when you try to put a piece of technology physically on top of the brain, even assuming that the procedure has perfect safety. Removing a Facebook-style, non-invasive neural electronics controller is as simple as taking off a hat; surgically removing permanent neural electrodes like those designed by Kennedy, or like neural lace itself, isn’t so easy.
They’re designed so that the brain can grow around them and integrate them safely. “You don’t want to take those out,” Kennedy says, “because you’ll rip out that part of the brain.”
The permanence of coming neural implants is worrying for a number of reasons, but foremost among them is the fact that science currently has a much, much better ability to pick up brain information than it does to make sense of it. An electrode implanted today has certain capabilities that are capped by a fundamental understanding of the brain’s activity, more than by the electrode’s ability to read and report that activity to a computer. Users might get an implant in 2020, fully comfortable with what it can do at that time, only to find that advances in brain science crop up years later to augment its existing mind-reading and -controlling abilities beyond what the user finds acceptable.
These sorts of risks are easily worthwhile for a paralyzed person whose basic mobility may be restored through an irreversible brain procedure, but is the ability to control a computer interface a good enough justification, as well?
Asked how a neural electrode could be removed or inactivated after insertion, Kennedy quickly argued that it’s hard to believe anyone would ever be so foolish as to request such a thing, but then said that if they did, the specifics would come down to the technology in question. In particular, the method of control is an important factor; keeping a neural implant secure, or even just ensuring it remains inactive after being shut down, is a matter of either stopping or protecting the streams of data moving both in and out of the implant.
The most secure, but by far the least practical solution would be a physical interface like a fiber-optic port. Kennedy, having conducted those sorts of experiments with neural electrodes in rodents, is intimately familiar with the potential drawbacks of a physical interface between the brain and the outside world. “Anything where you have wires coming through the skin,” he says, “is not a permanent solution by any stretch of the imagination.”
That’s consistent with the few details that are clear about oncoming neural devices. The aforementioned neural lace is actually only one of several technologies being surveyed by Elon Musk’s brain machine interface startup, NeuraLink. Though Musk’s venture could eventually decide to go with a different electrode technology, the neural lace approach does seem similar to DARPA’s plan to wire the brains of soldiers with electrodes that “could be delivered through minimally invasive injection.”
While we by no means know for sure that NeuraLink or any other player will settle on neural lace as its final platform for brain reading and manipulation, any project announced with an explicit plan for human applications is headed for some form of wireless control scheme.
So now, both removal of the implant itself and removal of its physical interface are no longer available options. How would someone deactivate a neural implant if they wanted to? More to the point, how would a person ever feel secure that it was staying inactive, when a re-activation even years down the line might be as little as a password guess away?
Kennedy thinks it might be better to specifically deactivate the communications and/or power portions of a neural implant to disable it permanently, though that would require a second surgery; if the implant was injected like neural lace, deactivation of the implant would be many times more invasive than its insertion.
Another option would be continuous blocking of the implant’s signals (it should be possible, since intracranial transmissions will have to be low-powered for safety reasons), but at that point we’re almost literally talking about a lifetime of wearing a tinfoil hat.
To say the least, getting even a perfectly safe neural implant is a major, life-changing decision, but that’s always been acceptable due to the seriousness of the issues these implants were meant to address. As Kennedy says, nobody is ever going to want their epilepsy to come back. Today, though, NeuraLink exclusively talks about its medical applications as stepping stones to its broader consumer ambitions. The human race will need the tech to roll-out on that sort of Muskian scale if it’s ever to realize the incredible bionic future that brain-machine interfaces could make possible — it’s just that a whole lot of early adopters could find that their lives have been irreversibly affected, along the way.
Asked about the idea of designing neural implants to safely degrade at a certain rate, thus making them inherently temporary, Kennedy responded, “It’s a terrible idea! Come on, that’s just ridiculous… We can make electrodes permanent, now. We should be focused on that.”
Not everyone is as adventurous as risk-taking scientists and entrepreneurs like Phil Kennedy and Elon Musk, however. These devices will only get approval for their very first rounds of human testing due to their proposed medical applications, and so it’s the idea of medical miracles that is setting the tone for the conversation. That tone is not changing as the conversation shifts toward more consumer applications.
Don’t worry too much, though. There’s plenty of time left between now and the first neural implant approved for wide consumer sale. There’s plenty of time for a sober look at the more nuanced implications of these technologies.
You know, the subtle stuff. Like how to turn the damned things off.