In the 2020 video game Cyberpunk 2077, transhumanism is commonplace. The game envisions a dystopic, sci-fi reality 50 years in the future, where high-tech gadgets are spliced and sutured into the human body to enhance abilities, augment performance, or replace disability.
As little as five years ago, Cyberpunk’s cyberware body modification came across as pure science fiction. Today, becoming the bionic man (or woman) is closer than we realize, and so are the potential problems that come with it.
“This is not a world where we have the opportunity to fail first.”
Addressing those problems, and how doing so might prevent real life from imitating Cyberpunk, was up for discussion at this year’s Upper Bound AI conference, in a seminar dubbed How to Avoid a Cyberpunk Dystopia while Advancing an AI-Powered Internet of Bodies.
The seminar was delivered by three University of Alberta (U of A) academics from the departments of computing science and medicine. The trio has been working collaboratively to research the intersection of digital security and human-machine interfacing in a project called Patient Centered Design of Reliable and Trustworthy Artificial-Intelligence-Enabled Prostheses.
Bailey Kacsmar is an assistant professor at the U of A and an Alberta Machine Intelligence Institute (Amii) fellow. Kacsmar leads research at the PUPS Lab, where she is assisted by Ph.D student Afari Darfoor. PUPS’ research focuses on digital privacy, security, machine learning, and human-computer interaction.
Patrick Pilarski is a professor with the U of A’s faculty of medicine, an Amii fellow, CIFAR AI Chair, and co-leads the BLINC Lab, a research organization studying, in part, the integration of advanced prosthetics and robotic systems.
Darfoor, whose thesis project focuses specifically on privacy and security for prosthetic devices—a field he’s calling idiobionics—laid out exactly what the “Internet of Bodies” actually is.
The next iteration of the internet
According to Darfoor, the term “Internet of Bodies” was coined in a 2014 Vice article by Meghan Neal.
“She goes into depth about how the next generation of the internet could be people themselves,” Darfoor said. “Over a decade later, we’re seeing that … materialize where we have this terminology of an internet of Bodies encapsulating a range and variety of technologies that are intimately and deeply integrated with the human body.”
While not quite the “gorilla arms” or “spring joints” available in Cyberpunk, Darfoor is talking about things like cochlear implants or bionic limb prosthetics. Those devices are not novel, but the last 20 years have seen huge technological strides not just in how connected they are to the human body, but also to the wider world.
“Over the years, we’ve started to see the clinical standard change. There have been amazing advances in actuaries in robotic technologies, in the sensors we can deploy into body-attach, wearable systems and computing technologies. We can now start to do pretty miraculous acts of machine learning and adaptation,” Pilarski said of the changing medtech landscape.
That might look like a cochlear implant that’s Bluetooth-connected to your Spotify playlist, enabling direct-to-ear hearing without AirPods, or AI imbued to automatically reduce ambient noise levels. It might also look like a prosthetic arm that leverages machine learning to perform complex, dexterous activities, all programmable from an offboard device.
Pilarski says as the field has changed, more advanced tech has crept into all types of prostheses.
“It’s very natural for designers and engineers to say: if we just added more sensors or technological capacity … these devices would be able to better monitor what people want, to be able to act on that, to be able to remember what they’ve done in the past, and to send information offboard of the device, to say a person’s cellphone,” she said.
That level of interconnectedness can provide a lot of benefit and convenience, but it can also expose some pretty large vulnerabilities, according to Kacsmar and Darfoor. Think of the kinds of identity theft, ransomware attacks, and spyware that are commonplace today, but applied to a body part or aid.
“We have this terminology of an Internet of Bodies encapsulating a range and variety of technologies that are intimately and deeply integrated with the human body.”
“If you remember the early Internet of Things, it had terrible security. You had people around the world turning on other people’s lightbulbs, exploding their rice cookers,” Kacsmar said. “Obviously, we don’t want to reach that point with [the Internet of Bodies]. If something happens to [these devices], you’re not going to be able to just use a different device or not use it.”
And then, of course, there’s the question that prompted Kacsmar and Pilarski to collaborate in the first place: what if a bad actor could take over your prosthetic and punch you in the face?
“We’re not yet there at the stage where [people] are able to figure out how to infiltrate someone’s arm and control it to punch them,” said Darfoor.
But Kacsmar noted that as AI integration becomes more commonplace in prosthetics, attacks such as data or signal poisoning, which could inhibit the usability of a prosthetic, could become more of a reality. It’s those types of potential outcomes that Darfoor, Pilarski, and Kacsmar are hoping their research might inform and mitigate.
“What we’re trying to do here is stop this from happening, from escalating to a point where even as we’re entering the phase of brain-computer interfaces … we can be cognizant of the security and privacy risks that come with them,” Darfoor said.
“This is not a world where we have the opportunity to fail first,” Kascmar added. “We need to be designing from stage one. How do we ensure that these devices have all these amazing advances … without introducing these risks, and that’s where this line of research is moving with what we’re doing.”
BetaKit’s Prairies reporting is funded in part by YEGAF, a not-for-profit dedicated to amplifying business stories in Alberta.
Feature image courtesy Jesse Cole for BetaKit.