Standards Gap in Personal-Care Robotics Highlights a Bigger Problem

Technological progress is a given, continuous and inevitable. From our privileged vantage point in the present, we have a majestic view of how technology has progressed through the valleys and plateaus of preceding generations. If we look more closely, however, we can see the progress didn’t traverse a direct, smooth trail. Looking back, we can see twists, turns, detours, and switchbacks, a journey replete with recalibrations and the need to overcome temporary roadblocks.

We can see now that we’re advancing into the uncharted territory of intelligent robotics, powered by increasingly capable AI. Just as in the past, though, the ascent is unlikely to be untroubled. We (I’m using the royal we to denote humanity) only gain a profound understanding of the future when it becomes the past.

Humanity learned a lot about both aeronautics and computing by studying and learning from malfunctions, instances in the realm of the former carrying more risk to life and limb than those in the latter. Similarly, we can expect that the learning curves associated with some categories of robotics will be more eventful and perilous than others.

One of those ares involves home (personal-care) robotics. You don’t need an overactive imagination to envision a spectrum of unfortunate mishaps that could ensue within the context of homes cohabited by humans, robotics, pets, and furniture. The mayhem quotient rises when one also considers the multiplicity of interactions that could conceivably occur among and between those elements.

You might reasonably object here, citing the existence and good works of standards bodies, such as the International Standards Organization (ISO). Indeed, you’ll be gratified to learn that the ISO is updating its 12-year-old safety requirements for personal-care robots. The problem is, robotics technology is moving fast, perhaps too fast for technically proficient and otherwise responsible engineers to fully appreciate the unpredictable dynamics between non-engineer human users and robots.

Robots and Humans: What Could Go Wrong?

In the virtual pages of IEEE Spectrum, this particular complication is emphasized in an interview with technology policy researcher Jae-Seong Lee, of the Electronics and Telecommunications Research Institute in Daejoeon, South Korea.

Lee defines the engineering problem in the following excerpt from the article:

“It is not simply whether a robot can avoid collisions or detect people in its path. The harder problem is that human-robot interaction is bidirectional. The robot changes what the human does, and the human changes what the robot perceives and does next. In other words, safety is not a fixed property of the machine alone; it emerges from the relationship.”

While engineers rigorously assess robot functionality, Lee explains, ISO 13482 “stops short of binding compliance criteria, test methods, or enforcement mechanisms for the hazards produced by the human-robot relationship.” What’s also missing is direct input from the types of humans who would be interacting with and using personal-care robots. As Lee says: “A robot that is safe only in standardized rooms, with healthy adults, under well-defined conditions is not really a domestic humanoid at all.”

What this situation highlights, as Lee notes later in the interview, is a governance gap in ISO’s methodological processes, a potentially serious oversight. The gap can be closed, of course, but it must first be recognized as a problem.

The Dilemma Persists, the Stakes Grow

In reading the IEEE Spectrum interview, I experienced a strong sense a déjà vu. Having we experienced similar difficulties before, in historical standards bodies populated by engineers from earlier generations? One could accurately state that computing and networking standards bodies in the pre-Internet era were technically astute but relatively unenlightened about how mere mortals would use and interact with personal-computing technologies. The result was an iterative, increasingly urgent downstream focus on ergonomics and the “user-friendliness” of technologies, a development that began in the early 1980s and gained momentum later that decade, thanks to Apple and others.

Interestingly, BYTE magazine — now long defunct — took instant offense in 1982 to the term “user-friendly,” as this quote attests: “The term "user friendly" must surely rate as the inanity of the decade. When was the last time you thought of a tool as "friendly"? "Usable" and "useful" are the appropriate operative terms.”

In good conscience, even now, I cannot take issue with BYTE’s verdict. Still, the resounding popular resonance and longevity of the term “user-friendly” attests to how unfriendly, even hostile, computing technologies were perceived by curious laymen who wanted them to be more accessible.

Personal computing back in the era of disco music could perhaps afford the cost and inconvenience of alienating and frustrating mainstream adopters. Nearly half a century later, as the pace of technological progress relentlessly accelerates, the stakes are higher. People and intelligent machines are interacting, and such interaction is likely to increase in complexity and scope in the years ahead.

Unfortunately, engineers and the standards bodies on which they serve probably cannot afford to succumb to the same blind spots that afflicted their technological forebears. A dysfunctional (in all respects) personal-care robot could wreak havoc, resulting in property damage, personal injury, and — in extreme but imaginable scenarios — death. An uncooperative personal computer circa 1981, disconnected from the world and relatively harmless, was merely a nuisance.

It’s a paradox of our time, but the need for standards-body vigilance and regulatory scrutiny is growing even as the market and political will to accept such responsibility is weakening.