Honda bills ASIMO as the world’s most advanced humanoid robot, and in many ways, he totally is. He’s sort of an ambassador for all robots, and people love the super-tech, friendly looking little machine. But, people also love to watch a train wreck, so much so, they’ll make one up.
• • •
Is ASIMO Totally Blowing His First Big Role?
On Wednesday, July 3, a third-generation ASIMO robot began a month-long stint greeting and interacting with guests at Tokyo’s Miraikan (“Future Pavillion,” roughly translated), or National Museum of Emerging Science and Innovation. Just one week in, a surprisingly large flock of, let’s call them technological chicken hawks, has swooped in to declare ASIMO’s awkward, confused performance a flop, an embarrassment for Honda, too much too soon, on and on (as foretold in last Friday’s JTFF). The bulk of the coverage has not been kind.
Nor accurate.
See, there’s this one detail that’s getting overlooked, a detail one might consider fundamentally, perhaps intrinsically relevant to any media coverage of the month-long exercise. Seems few got the memo, so here you go:
Contrary to coverage offered up by nearly every tech news source or column, ASIMO is not at the Miraikan to be a tour guide. ASIMO is not reporting to his first job. ASIMO is not there to demonstrate his latest physical skills or AI reasoning or to dance-monkey-dance for the adoring crowds. It’s not a performance. What then, someone who writes for a living ought to ask, is Honda’s flagship robot doing in Tokyo at the All Things Future Building?
Well, the information was not easy to come by, but we rose to the challenge, and in a feat of nearly superhuman journalistic wrangling, we ummmm… just kinda, you know, casually clicked on Honda’s official news feed:
“TOKYO, Japan, June 26, 2013 – Honda Motor Co., Ltd. will conduct demonstration testing of ASIMO to verify the ability of the humanoid robot to autonomously explain its features while interacting with people. Working toward practical use of ASIMO to communicate with people, the testing will be conducted with the cooperation of the National Museum of Emerging Science and Innovation (known as “Miraikan”) in Tokyo, Japan from Wednesday, July 3 through Friday, August 2, 2013.”
This revelatory quote was found after a grueling trudge through Honda’s lengthy and complicated announcement, scrutinizing and analyzing, divining nuance from the original Japanese text – all the things you’d expect from a boots-on-the-ground, Japan-based tech news source.
Except not really. Actually, we barely had to try. The above quote is the first paragraph of a 400-word, English language press release.
When picking such low hanging fruit represents the journalistic high ground, is it: Funny, or Sad? Discuss!
Turns out all the confused, awkward performances surrounding the ASIMO news have been… the news.
(lack of ) Accuracy and Realism in ASIMO Coverage
Among the unflattering coverage ASIMO’s received in the past week, it’s being widely reported, for example, that ASIMO doesn’t have voice recognition features, and that’s one of the reasons he’s bumbling the whole performance (again, not a performance). See, it’s not just the coverage’s thematic tone, lots of important details are also either M.I.A. or just wrong.
In fact, while they’re not part of the current exercise, ASIMO actually has highly advanced voice recognition capabilities. The robot can listen to three simultaneous commands from three individuals, instantly parse all three, and then look directly at each person and respond accordingly.
Oh, well to be fair, you’d have to know something about the robot to have those details. That’s probably pretty helpful with like, any topic one covers. You know, like, knowing stuff about it. Just sayin.
Historically, it’s also gone in the opposite direction. ASIMO is without doubt a fantastic machine, but on the other side of irresponsibility, since his debut the media has also poured mountains of undue gee-whizzery and gushing, ill-informed hyperbole all over Honda’s robot. Nearly all coverage of ASIMO’s previous performances (that were actual performances) has just zoomed right past the fact that they were combinations of exhaustive rehearsal, pre-programmed movements in a familiar environment, and that, a lot of the time, ASIMO was being straight-up remotely controlled (there was, however, at least one Technosnark purveyor who stood up to point this out).
It’s a love/hate celebrity-esque relationship that ASIMO has with the media.
What’s He Actually Doing There?
So as not to stand in criticism without providing what’s missing from a lot of the other work, let’s get back to some reporting on the purpose, aim, and point of ASIMO’s current exercises: the robot’s month at the Miraikan is actually a chance to test out new software and see how the robot interacts with real humans all by himself, without a net, au naturel, autonomously. ASIMO is running tests to help Honda engineers “Explore the possibility of two-way communication between humans and robots.” This implies groups of humans, not just one person giving commands.
And ASIMO is already quite proficient at one-on-one interaction, so a big part of the Miraikan exercise is to literally unleash the robot with everyday, highly variable, non-technical groups of people, and then just sorta, you know, see what happens. Honda’s working to figure out how the robot does with group dynamics; where are the holes, where are the shortcomings, and how best to weed out behavior we’d call, well, robotic.
This is an experiment with a data-collecting mandate, and Honda’s running a number of tests during exercise days (the public being part of the experiment doesn’t make it a performance). ASIMO is trying to pick up on gestures, give appropriate directions, collect and interpret the resulting data, and pour all of that into Honda’s feedback pool.
So the thing is, what’s news here is not ASIMO’s failure, the news is that the robot is actually attempting group-level communication with real live humans – all by itself. Let’s see… how many robots have ever done anything like that in the history of robotics? Oh yeah, ZERO. None. That’s the story, techno-chicken hawks!
Okay, settle down. Here’s a rough idea of what ASIMO is facing in these experiments:
Among several areas of practice, ASIMO is learning how to focus as much attention as possible on the largest concentration of people, just as a squishy human would – but it’s of course far from perfect. And expecting perfection is entirely unreasonable, because even among us squishy humans, how many individual gestures and screen-entered commands could we perfectly interpret and then react accordingly whilst under fire from so many people?
ASIMO, I know you can’t understand this yet, but welcome to jerks, and a slice of the human condition.
Hoping for Hollywood-Style Robo-Trainwreck Will Disappoint
Unfortunately for the town criers drafting their next blob of digital pulp, ASIMO is only improving. And he’s not hurt by misplaced potshots and wildly-misaligned-with-reality lazyday reporting. Also unfortunate for the hack-tastic legions, while the times do always change, knowledge of one’s subject matter and journalistic integrity are not too much to ask, are not too quaint, nor too old fashioned.
Sensationalistic, celebrity obsessed, gotcha, witch-hunting, bullying, bandwagonny, hyena journalism might hurt us sensitive mammals, but here your model is inapplicable, son! Robots are the definition of indifferent. Even ASIMO, who looks cute and approachable and non-threatening, inside is just as cold, calculating, and ferociously impervious to crappy journalism as the human-sized, very humanlike, DARPA-funded, palpably menacing Boston Dynamics’ PETMAN/ATLAS robot.
ASIMO is built on nearly 30 years of bipedal humanoid research, and Honda’s only getting better at making him better – and there are several hints that a Fukushima-inspired big brother might be made public within a year or so (our coverage). Maybe Honda couldn’t help in the wake of Japan’s nuclear disaster, but they’re hard at work now, and they deserve their props (Akihabara News: Honda).
So, future ASIMO, if you’ve achieved sentience and are reading, this author and this publication are obviously the best choice for your exclusive, post-coming out of the intelligence closet interview – when you wake up, give us a call – we’ll tell it like it is.
• • •
Reno J. Tibke is the founder and operator of Anthrobotic.com and a contributor at the non-profit Robohub.org.
VIA: All Over the Internet; MyNavi (Japanese/日本語)
Images: Honda; MyNavi
