CorTec and the human/AI fusion
A human and a computer walk into a chat room…
Tl;dr: in this post, I’ll discuss a neurotechnology company called CorTec and their quest for human/AI fusion, then offer some thoughts on the latter.
At Society for Neuroscience 2017, Ladan Jiracek of the Neural Implant Podcast interviewed Dr. Jörn Rickert, the CEO of CorTec. CorTec is a six-year-old company that designs and produces neural implants; they’re based in Germany, and their technology builds upon a decade of foundational university research and development — i.e., they’re academics who’ve taken their expertise and brought it to industry.
CorTec has spent years working on the CorTec Brain Interchange, a closed-loop, all-in-one neural interface whose creation pursues the lofty and intriguing goal of fusing humans with artificial intelligence.
The Episode
First up: let’s dive into the technology that CorTec has been developing! Currently, the company specializes in and sells soft, configurable electrocortigraphic (ECoG) electrode arrays. ECoG is a specific type of intracranial (within the skull) neural recording device that sits on top of the brain’s cortex (which is the outer-most layer). The notion that it sits atop the cortex instead of within the cortex or deeper brain structures is important! Why? Because it comes with tradeoffs. ECoG is less invasive than other implants because it doesn’t penetrate brain tissue; however, it can only record from larger populations of neurons at the surface of the cortex.
For their upcoming innovative product, the CorTec Brain Interchange, CorTec has chosen a specific ECoG configuration. Below is an image of the entire flattened implant, and an image of it disassembled into parts.
With the CorTec Brain Interchange, the company is moving into a more advanced, and exciting, product area: they’re incorporating this electrode array into a system capable of recording from the brain intracranially, outputting that signal to a wearable computer which can decode neural information, and stimulating the brain based on the recorded information — all in realtime. Let’s abbreviate “CorTec Brain Interchange” with “CBI,” to preserve sanity.
The CBI has three components: a soft ECoG array with 32 channels (each channel can record or stimulate neural activity), an “internal control unit” that receives signals from the ECoG array (“internal” because it sits within the patient’s skull), and a small externally visible unit attached to the head that a) inductively powers the internal control unit, and b) communicates with a computer pack that performs computations based on the ECoG recordings.
As part of their research/development and for regulatory reasons, CorTec first has to test these implantable devices in animals before they can be used in humans. Jörn commented that CorTec has been able to leave the ECoG array in a sheep for ~1.5 years to date, and impressively, when the ECoG array was removed, the device came out looking like new. CorTec will be pursuing human trials of their product in 2018, pending European regulatory approval.
As a quick aside: sheep are frequently used to test neural implants. They’re intelligent animals with large brains, and have primate-like cortical folds, in addition to primate-like deep brain structures (i.e., the basal ganglia). Therefore, they provide a way to understand the safety and effectiveness of a neural implant before implanting into a human and risking a human life. Here’s a visual comparison of brains from different species — note the similarities! If you want to read more about the relationships between different species’ brains, check out this paper.
So, returning from our sheep brain detour…what’s special about the Brain Interchange? As I understand it, it’s a combination of two things: ergonomics and algorithms. In terms of ergonomics: the CBI brings together a number of important technological components into one simple package. They have a biocompatible electrode array (i.e., the brain doesn’t try too hard to reject the array, the process of which damages electrodes and reduces their ability to record/stimulate), an inductively powered controller (this is important because patients with this implant won’t have to have a battery pack implanted), the ability to stimulate and record from the same electrodes (this is important for applications like controlling a prosthetic limb, where a patient needs to be able to control the limb and also feel what the limb is touching), and a conveniently attached computer pack. To put this intuitively, if CorTec ever decides to advertise via infomercials, they’ll probably use the phrase “All in one!” (for only three payments of $19.99, etc.).
The computer pack, in my view, is the most compelling part — and also goes the longest way towards achieving CorTec’s mission of AI/human fusion (which, as I’ll reflect on below, bears tremendous ethical and political implications). Traditionally, in order to interpret neural signals and decide how to stimulate the brain in response, researchers need a large equipment array in a laboratory setting. Currently, this makes it implausible for patients with neural implants to leave the laboratory and access the implants’ therapeutic value in their daily lives. CorTec’s wearable control system (similar in aesthetic to the wearable computer being developed by mixed reality company Magic Leap) opens up the opportunities for a patient to take their therapeutic benefits out of the laboratory.
And then there’s the artificial intelligence component: sitting in this wearable computer are a series of algorithms that help interpret complex neural signals recorded by the ECoG array, and thereby decide the best way to stimulate the brain. As articulated on their website, these algorithms make the CorTec Brain Interchange a “responsive system,” so that the “therapy [can be] adjusted to [the] current need” of the patient. In other words: instead of indiscriminately stimulating the brain based on a best-guess of a patient’s symptoms, these algorithms intelligently decide when to intervene and when to leave the brain to its own devices (get it? devices? medical devices?).
This application of artificial intelligence (/machine learning) algorithms is far from dystopic science fiction in that it has limited functionality and real therapeutic value. It’s clear, however, that Jörn and his team at CorTec are thinking about future-facing implications of their work. Jörn reflects:
“As Elon Musk said, we need to probably enhance ourselves to be able to compete with AI, and not become outmatched by them…and that’s essentially a political debate…Personally, I would say that — at least what we had in mind when we started CorTec — it’s about helping patients.”
So, let’s take this cue to dive into the topic of human enhancement.
Ruminations
The idea of enhancing the human brain, especially by fusing it with artificial intelligence, has existed in the annals of science fiction and science reality for a long time. We see it in non-implantable forms, such as with Iron-Man-suit-wearing Tony Stark and his AI, Jarvis. Tony’s physical abilities are of course augmented by his suit, but he frequently tells Jarvis to handle complex tasks he can’t focus on in the moment, thereby “enhancing” his intellect (which, frankly, is already pretty extraordinary).
We also see human + AI as an inspiration for current real-life neural interface work — as Jörn alludes to, the motivation behind Elon Musk’s founding of the BMI company Neuralink can be stated as, “our best bet as a society of not being overtaken by a runaway artificial intelligence is to become a biological artificial intelligence ourselves.” Crazy, crazy, stuff. I’ll refrain from offering my opinion on the topic of the dangers of AI here, in part because I’m still developing it.
Human enhancement in general, of which human/AI fusion is a salient form, falls firmly within the bucket of “augmentation,” as I define it here. It requires fierce ethical rigor during its development and application, and I agree with Jörn in that these ethical considerations ought to be discussed by the public, and accessible to the public.
Have humans ever faced a similar scenario of choosing so overtly to change themselves? In a sense, we frequently elect to change ourselves in aesthetic domains: impermanently, we wear clothing, cut our hair, wear colored contact lenses; permanently, we receive tattoos and piercings.
Another method of changing ourselves is inventing and using technology. Illustrative anecdata: “life is so much different with smartphones” -any person who’s been alive for long enough to remember what life was like pre-smartphone.
More heavily, we can also consider the burgeoning field of gene therapy, in which diseases are treated or prevented by modifying our DNA. This field brings with it intense ethical questions, some of which are outlined here. As an interesting side-note, in his book Superintelligence, Nick Bostrom offers back-of-the-napkin calculations for the impact of orchestrating genetic selection for intelligence. Make no mistake, it’s a remarkably hairy topic.
With respect to aesthetics, general technological innovation, and even gene therapy: I would argue that the idea of plugging a computer into the brain and using this to incorporate artificial intelligence into subconscious or conscious processes is on a whole different plane. Why? Because we’d be controlling the substrate of human experience with precision, as opposed to grossly affecting the inputs to this substrate from the outside world (or, in the case of gene therapy, changing the genome that develops the substrate in an environmentally-dependent way).
Given the weight of brain-based human enhancement, we clearly need this to be a topic discussed in the public forum; i.e., politically. I’m not sure yet how I think this can/ought to occur — I’m working on these thoughts, and would love to hear anyone’s suggestions. As part of my exploration on this front, I have some planned readings in neurophilosophy and neuroethics. Thoughts-in-progress:
- Without conferring any value judgement here, people will be magnificently resistant to brain-based augmentation. My dad, as a moderately change-resistant individual, doesn’t like getting new cellphones. I’ll let you know how it goes when I pitch him on a new brain.
- The naturalistic fallacy/appeal to nature is going to show up. I’m not an expert, but this is basically the notion that “because nature created something this way, it ought to be this way.”
- Security concerns are real. The collection of behavior data by companies like Facebook is already fiercely debated; brain data will be maybe an order of magnitude more polarizing (Jörn talks about this in the last quarter of the podcast episode).
- Brain-computer interfaces that enable human/AI fusion will be extremely expensive. Which means they’ll initially be accessible only to those with money + power. Which means the powerful become more powerful…(Tony Stark, with his aforementioned Jarvis augmentation, is a fictional example I find intriguing).
At some point down the road, as I coalesce perspectives on how to ethically and democratically introduce human augmentation through brain technologies, I’ll write an extensive post dedicated to the topic (although I won’t promise you when this will be, yet). Regardless, I’d be remiss not to offer the significant disclaimer that these questions are probably a decade (if not more) from reaching public prominence, and it would be an awful mistake to let them overshadow the incredible therapeutic benefits of brain computer interfaces that the hard work of folks like Dr. Jörn Rickert and the CorTec team will bring to patients in the near future.
In Conclusion
In this post, I talked about the Neural Implant Podcast interview with Dr. Jörn Rickert of CorTec, a company building a closed-loop implantable neural interface for therapeutic applications of human/AI fusion. Inspired by CorTec’s mission, I offered some nascent ruminations on the human/AI fusion.
One of my primary goals with The Substrate is to encourage an ethics-first conversation about brain-computer interfaces. To that end: comment, write me at hello@thesubstrate.com, or find me on Twitter (@averybedows). I will respond! (seriously, I will).
Until next time,
Avery
