Neuralink is a Safety Critical System
I do recognize the awesome potential to be realized with these developments but as with all great advances there will be good and bad to come of it. Also, don’t mistake my concern as wild paranoia of the future. My objective here is to think broadly… even if it does come off a little crazy. Seriously though, this IS pretty crazy.
Safety First
Early in my career I was assigned to assess the criticality of weapon system software. It was during this time that I developed the skill of envisioning that which has yet to happen. I spent a lot of time imagining the occurrence of various operational and non-operation system scenarios with the intent of identifying unsafe conditions (inadvertent launch, fratricide, etc.). Safety analysis is pessimistic by nature.
So of course when Nueralinks July press-conference had concluded my gears were turning. My initial focus was on the interface. Not so much the physical interface but more so the data interface and what the data may represent. Will the data be protected? Will the ability to precisely receive brain activity eventually lead to the ability to precisely write back. Could written information eventually be characterized as a command? Would timing matter? What could result from erroneous information?
Neuralink is in its infancy but these questions are important. As with any safety critical system the main focus up front must be Safety by Design.
Rest assured, Neuralink’s current implementation is not bi-directional. Meaning data will not flow both ways. Today’s version of the system simply attempts to read brain activity and interpret it to control connected devices like cell phones, computer mice, and remote controls. The goal is to help those suffering from decreased motor functions.
Protecting Personal Data
The foundational technology behind companies like Neuralink is their ability to interpret the meaning of your brainwaves. While they advertise that this information will be used as a means to control devices like the before mentioned remote controllers it wouldn’t be far reaching to assume that they also infer additional meaning used in other ways. For instance — the masses have recognized the predictive power of big data. Will real-time monitoring and interpretation of brain activity wield even greater control over civilization?
Of course, this begs the questions — How will this kind of personal data be protected? What if the interpretation of brainwaves suggests illicit activity? Will this eventually lead to a new era of predictive policing? Think Minority Report. Will there be “for profit” attempts to use the data collected to develop predictive models?
Hollywood Got There First
Ok, now for the crazy talk. I can’t help but think of (recall hehe) the film Total Recall. Not the reboot. Remember the scene where Arnold’s character Douglas Quaid visits Rekall? They’re the company that planted false memories into people’s brains allowing them to experience the thrill of Mars without having to travel there. Well, you may also remember that there was a hiccup with the process and a false memory was actually implanted in Quaids mind. As no surprise, Hollywood made sure Athings didn’t go well from there.
During this scene of the film Rekall is treating the brain no differently than we treat any old ordinary piece of physical memory. Something like a hard drive. Neuralinks technology is evidence we are heading down the same path (as far off as it may be). In the case of Rekall the brain is an organic storage device and in the same way that we can read/write a hard drive we can rest assured that companies like Neuralink will attempt to and one day succeed in developing techniques to read/write the brain. Buyer Beware!
A Fleeting Memory
When we accidentally wipe out a hard drive it gets reformatted. What happens if a brain is wiped out? Or some portion of it? How can we safe-guard against this? Will it be possible to backup portions of the brain? Some lessons may be learned from the IT industry.
We Need Policy
Understanding and mimicking the approaches IT professionals take to protect storage devices and personal data against erroneous writes is a good place to start. The key (and point I’m trying to raise) is publicly establishing a plan and building a framework for future safety mitigations in this growing field of brain implant technology. Let’s start now.
About Me
By profession I am a system safety engineer. As an Applied Mathematics graduate student my research interests are Safety and Security Analysis Techniques and Design Methodologies for Complex Systems, the Safety of Artificially Intelligent and Autonomous Systems, and Systems Theory.
