Robot Reality Check: The Truth About Ebola Bots And Iron Man Suits
Imagine you’re in Guinea, ground zero for the Ebola epidemic. Your country is struggling to provide basic medical services for...
Imagine you’re in Guinea, ground zero for the Ebola epidemic. Your country is struggling to provide basic medical services for thousands of sick and dying. Corpse disposal is a full-blown crisis. The World Health Organization is making desperate pleas for more money, and more personnel, to help contain this plague. You happen to be online, and on a site called Defense One, reading a story entitled, “Here Are Some Robots We Could Use to Fight Ebola in Africa.”
Now this is news you can use. This handy list of robots waiting to be deployed includes iRobot’s Ava 500 telepresence bot, a tablet-headed machine marketed primarily as a tool for corporate executives, allowing them to remotely roam the office and attend meetings from across the country or world. This robot costs $69,500. Defense One’s story also talks about the Japanese Robokiyu Rescue Robot, an experimental one-off bot built for the Tokyo Fire Department in 2007, and never used for any real-life rescue operations. It slowly, not-so-impressively hauled mannequins from one spot to another. Maybe this robot could haul away infected bodies, to be disposed of by other incredibly expensive robots. Who knows? When you’re covering an epidemic by writing about robots designed to improve the teleconference experience for employees at Fortune 500 companies, and robots that are pure, YouTube-fodder vaporware, anything can happen!
I’ve written before about the various reasons that smart people seem intent on saying and writing dumb things about robots. Often it’s a matter of limited knowledge of complex technology, a gap in understanding that even technology writers tend to fill with credulous references to long-dead projects (like the Robokiyu). Sometimes, it’s just bad speculative fiction disguised as journalism, like when reporters jumped on the hype train with Honda, and imagined the Asimo humanoid being swiftly upgraded to join the recovery effort at the Fukushima power plant. Asimo never did make it to the disaster zone, and neither did a Japanese exoskeleton called HAL. Robots rarely have to be useful or feasible to justify coverage. They only have to be robots.
Ebola-fighting bots are the most recent, and probably worst example of this recurring journalistic failure. But misinformation loves company. Writers and editors are happily ignoring the technical and financial realities surrounding a wide range of robotic technologies, and assuming that all manner of advanced machines are on the cusp of deployment. Here are four classes of robotic devices whose estimate time of arrival has been greatly exaggerated.
Last month, Elon Musk told The Wall Street Journal that fully autonomous cars were just five or six years away. Though I wouldn’t want to spar with the SpaceX and Tesla CEO over the finer engineering details of those systems, Musk is echoing a growing sentiment among automakers, and the journalists who cover them. From Audi and BWM to Cadillac and Mercedes-Benz, the consensus is that robot cars are right around the corner.
As Lee Gomes pointed out in a recent (and excellent) Slate story about Google’s self-driving car:
“It can’t consistently handle coned-off road construction sites, and its video cameras can sometimes be blinded by the sun when trying to detect the color of a traffic signal. Because it can’t tell the difference between a big rock and a crumbled-up piece of newspaper, it will try to drive around both if it encounters either sitting in the middle of the road.”
These issue are the same ones that have confounded autonomous cars for years, all the way back to DARPA’s Urban Challenge in 2007. Even with redundant sensors—radar for long-range detection, and vision, lasers and ultrasound for closer-range perception—robotic cars can see phantom obstructions in the most optimal conditions, much less during a blizzard, or on roads whose 3D-mapped contours don’t line up with the detours and lane closures caused by temporary construction.
Also, keep in mind that it takes at least 3 years, if not a half-decade or longer, for a brand new automobile to be designed, tested and manufactured from the ground up. For fully autonomous cars, or even vehicles that can shift into a truly self-driving mode, to reach the market by 2020, they should already be in development, right now. And if that were the case, no amount of secrecy would prevent the inevitable leaks. There are half-measures and half-steps from the likes of Mercedes-Benz and Volvo, but nothing that makes the leap from advanced cruise control-like features to true autonomous driving. If it existed, in any form, we would know about the world’s first commercial robo-car. There’s no other evidence necessary that autonomous cars are further off than Musk, and much of the media, would like you to believe. Remember, it’s in the best interest of everyone involved to keep you, the customer, in a state of perpetual anticipation. But you can believe in the hope of robot cars, without buying into the near-term hype.
Iron Man Exoskeletons
Journalists appear to have one, and only one way into their coverage of real-life exoskeletons: A not-real superhero. This past May, a LiveScience story about the preliminary development of an armored exo suit for U.S. special operations personnel featured the usual intro:
“Move over, Tony Stark — the military could soon have its own ‘Iron Man’ suit, a robotic exoskeleton designed to augment human abilities on the battlefield.”
Along with being terrible writing, a notch below what you might hear on your local newscast, this boilerplate lead-in is blatantly misleading. As of May, the TALOS suit didn’t have a coherent design or set of goals. I spoke to Special Operations Command (SOCOM) at that time, in the interests of putting together a feature on the project. SOCOM’s hope was to unveil multiple versions of the suit this past summer. Naturally, I wanted a sneak peek at the three prototypes.
SOCOM politely declined, with good reason–there was nothing to see, and little to talk about. The three planned prototypes were going to be non-functioning concepts, the battlefield exo equivalent of automakers pulling the tarp off of a non-functioning concept car whose hood won’t open, because it hides an engine-less void. The designs were going to be props, designed to gauge public interest, and attract more companies to get involved with the program. TALOS was a mindset, more than anything, an elaborate request for proposals from robotics firms, as well as from contractors who might supply next-generation communications systems and ballistic armor. In fact, SOCOM told me that the program’s eventual success assumed that new battery and armor technologies would show up. How would special operators wind up in suits those augmented strength might offset the weight of additional armor? By hoping for the best. Forget TALOS as a mindset—it was a hope and a prayer.
I assure you, the point of all this isn’t to boast about my investigative prowess. It required a grand total of one e-mail and one phone call to discover just how ephemeral TALOS was. Now, months after the proposed debut of those prototypes has come and gone–they were reportedly reviewed by SOCOM, but not released publicly–the only sign of life in the TALOS program are a few stats provided to Popular Science about the current prototype’s weight. But even that mention is a footnote within a larger item about another exo device. It’s the right way to write about TALOS, in other words, with none of the blind credulity or dorky Iron Man references. On the whole, none of the skepticism or basic fact-finding that journalists might apply to other technological fields was present during the various cycles of TALOS coverage. Robotics gets a pass, precisely when it shouldn’t. When everyone’s gushing over the prospect of armed forces decked out like comic characters (minus the flying and the ray guns and the tank-like armor), who wants to be the grumpy outlier who grinds the collective geek-out to a halt?
Meanwhile, exoskeletons that don’t purport to augment ability, but rather work to restore it, have quietly transformed into legitimate medical devices in the United States. In June, ReWalk Robotics received the first FDA clearance for its Personal System exo to be used in the home, and out in the world. ReWalk has no plans to build military systems, or to provide superhuman strength or endurance, which might explain why its history-making approval received so little attention. It’s further evidence of the fundamental disconnect between the robotics stories that catch fire on Facebook and Twitter, and the ones that describe robots that are making a tangible, present-day impact. The bots that don’t exist will always be more exciting than the ones that are already among us.
Hunter Killer Killbots
In May, the United Nations held a summit to discuss the future of Lethal Autonomous Weapons (LAW), including the possibility of banning such systems. The press, naturally, had an extended field day, littering news feeds in the weeks before and after with headlines about killer robots. The summit didn’t lead to anything tangible, but as recently as August, it continued to supply journalists with opportunies for strenuous hand-wringing. The UK’s Telegraph quoted Angela Kane, the UN’s high representative of disarmament, as saying:
Kane equates autonomous cars, a technology that’s still an unknown number of years away from viable commercial deployment, with weapons that are free to attack targets, based on their own judgment. She also describes warfare as “becoming increasingly automated,” when what she really means (we can only assume), is that warfare is more robotic, since the many drones that are currently in the air are remotely piloted by humans.
Kane is nonsensical, in other words. She demonstrates a clear lack of knowledge or understanding about any branch of robotics, much less the systems that are her job to monitor. But the story plays along, and is comprised almost entirely of direct quotes from Kane, who also says, “I personally believe that there cannot be a weapon that can be fired without human intervention.” It’s not a very bold stand to make, considering that no one on the planet appears to want that sort of weapon. Even the much-discussed armed surveillance bots on the border between South and North Korea are theoretically capable of firing without intervention, but that feature has reportedly never been engaged, and appears to be more of a technical possibility than a valid feature. Strictly speaking, any armed unmanned system could be set to fire autonomously. But for whom would that be useful, or desirable?
I’ve argued in a previous post that the recurring furor over killbots is a sign of our collective cluelessness about robotics, but journalists are complicit in overselling a threat that exists only in science fiction. The Telegraph story finds its way back to reason in its final paragraphs, with an expert from IHS Jane’s Defense Weekly clarifying that there are no known LAW programs in development, and that commanders would see them as a liability, not an asset. Still, rants like Kane’s dominate LAW coverage, and their uninformed perspectives are largely repeated with little to no context or analysis.
Fun Robots That Fight Tragic Plagues
Despite the fact that it will only increase the traffic for a patently foolish post, I would urge you to read that Defense One piece about the Ebola-fighting bots, and to consider its immodest proposals.
How do you help impoverished African communities combat the worst Ebola outbreak in history? With robots designed for rich people. Send telepresence robots into overwhelmed hospitals and aid stations, despite the fact that local internet bandwidth is unlikely to provide solid connectivity for navigation, or high-resolution video feeds. And while bots like iRobot’s RP-Vita are designed for use in hospitals, no current telepresence robots are built to survive frequent and rigorous sterilization measures. Reconfiguring them to that end is a non-trivial problem, and would call for non-trivial additional funds, at a time when the Ebola response is already tragically short on resources.
The post also mentions the possibility of using OskKosh Defense’s TerraMax autonomous truck, for supply deliveries. This is the same model of robot truck that crashed into a mock house during the DARPA Urban Challenge, in a setting where GPS signals were plentiful, and every participating vehicle was loaded with a breadcrumb trail of GPS waypoints to keep them on the road, meaning location data that exceeded what a standard automobile might receive via satellite. Other self-driving bots in that race had technical difficulties, but the TerraMax slammed into a house. Now imagine sending that truck, or any other autonomous vehicle, for that matter, down poorly mapped dirt roads, filled with people, animals, and/or debris.
None of this technology is ready to ship to Africa, or appropriate for such a deployment. The impetus for the post was a blog post by Texas A&M roboticist Robin Murphy about an upcoming workshop on safety robotics for Ebola workers. Murphy, who is director of the university’s Center for Robot-Assisted Search and Rescue (CRASAR), suggests activities where robots might be able to help. But what she makes clear, and what the DefenseOne writer does not, is that the challenges are large, and in need of serious discussion. From Murphy’s follow-up post:
The TerraMax is a military vehicle. So is the iRobot Kobra 710 featured in the DefenseOne story (the company markets it for hazardous materials handling as well, but that’s an ancillary application for a bot that built and priced for the defense industry). Murphy throws water on the idea that getting robots into the Ebola fight is a matter of simply picking out existing models. Defense One does exactly that.
It’s clear why this sort of story gets written. You can almost see the gears turning, the pitched or assigned premise–people love robots and hate Ebola—self-assembling into a bouncy listicle that ignores the logistical realities of robots, as well as Ebola. It’s the perfect headline and topic to share, too. Ebola is utterly terrifying, and, like most major plagues, seems strangely impervious to modern technology. Which is why robots are the perfect tool to imagine using against it, a borderline imaginary technology that reporters continue to pretend to understand, when really they’re just regurgitating the results of hasty, context-free Google searches.
Robotics is an exciting, and relatively fast-moving field of research. But the most promising systems won’t be rushed into real-world duty by a few ignorant or impatient journalists. These stories don’t hasten the speed of innovation. They create expectations that can’t be met, and contribute to the general glut of misinformation about the state of robotics. And let’s not be coy—they lie, in the cheerful but knowing tradition of writers predicting a jetpack in every garage and a lunar vacation for every family. All too often, robotics stories are to technology journalism as Cosmopolitan’s sex tips are to relationship advice. They’re a joke, and not a particularly good one.