Skeptical reporter is rightly skeptical.
The demo was impressive, I’ll give them that. A machine - Google’s voice assistant - booked an appointment at a hair salon. And then a table at a restaurant. Not online, or through some automated system. But by talking to a human. Over the phone. Blimey.
Now. Let’s start with the obvious question. Does it work? We don’t know. Frustratingly, Google was unable to show us this technology in action. We have no idea if the calls shared today were the successful calls out of many, many attempts - nor do we know if the recipient was prepped beforehand. We don’t know how easy the system is to fool, or just confuse enough to render it useless. Anyone who uses Google’s Assistant today knows how often it stumbles over a lot of basic requests.
So, while others in attendance have referred to today’s demo as “stunning”, I’ll retreat to something more sensible: it’s promising.
As is increasingly the case, the key hurdle here may not be one of technological limitations, but societal. Start here: will the recipient know they are talking to a machine? Do they deserve to be told? Will Google monitor each call and learn from its contents? If so, how would the recipient consent to that - as the law, in many places, demands - ?
Assuming that gets figured out, think of it from the other side. As a human being in a service job, how would you feel about customers who couldn’t be bothered to call you themselves? Like a shop worker might recognise an irritating customer’s voice, staff will surely start to notice the telltale signs of an AI speaking to them. I’m not sure I’d give it the time of day. Then again - if it brings in a lot of business, workers might be told to just get over it.
Or maybe, the salon could get an AI assistant of its own. And then it's AI talking to AI. Booking appointments. Planning. Plotting. Waiting.
Oooh, sinister stuff ! Like Terminator only with more wine and cheese and... well, not really very similar at all, really. Only, surely what will happen is that restaurants will develop online booking systems (hint : better idea, get someone to develop an app for restaurants to do this - if this doesn't already exist then why the hell not ?) so they don't have to talk to robots. And there's no point getting an AI to automatically book one of them, except in that the user could ask an assistant to book a restaurant and it could automatically trawl databases to find the best match to stated preferences.
http://www.bbc.com/news/technology-44049904
I have robots calling me every day to ask me for money. I ask "are you a robot?" and sometimes they even have a response for that before they return to the regular script. I hang up on them without further comment.
ReplyDeleteNow. I might persist through a booking script if there were money/customers in it for me, but I'd much rather be emailed or have a booking system that people use. And doesn't that have to be simpler than programming your robot assistant to phone for you?
Ol' Dan has a few ideas about how this is all going to evolve. It's what I call Weese's Linguistic Plastic Sushi.
ReplyDeleteHave you ever been in a highly multicultural / multilingual situation, trying to buy anything? I am perpetually amazed at the linguistic resourcefulness of everyone involved, buyers, sellers, assistants of every description.
There's no real lingua franca involved, though there is a substantial chunk of the communication which must be done in some arbitrary language. It's whatever works in a given context, right, between ? Nothing terribly profound or insightful in any of that.
Let's posit an AI in record mode. It's not interpreting anything just yet, it's listening to a street vendor in operation. But more importantly, it's watching. It's learning the language of Plastic Sushi by listening to the customers and what the vendor says in response.
How long does it take the AI to learn the Plastic Sushi language ? You'd be surprised how quickly it can be done. The vocabulary is finite and plenty of what isn't significant can be ignored. Furthermore, every word can be directly mapped onto a known language.
But the interesting part will arise when both the customer and the vendor have AIs. They introduce their AIs to each other, both dump their lexicons and maps into each other, hey presto, both have the exact phrases needed for transacting business - within certain limits of course, but it would be a form of augmented speech.
...coincidentally I just had a human call me while I was posting the last comment. In a strong Indian accent they told me they were Mark Williams from Microsoft and they wanted to talk to me about Windows, how was I doing today?
ReplyDeleteI hung up on them, too. Turns out I have poor tolerance for scripts even when they're delivered by reasoning beings.
My guess is when these become convincing and more varied in they capabilities they will also be legislated to identify themselves as "agents" or bots. At least the ones that contact a person. The reason: Marketeers will use it to influence behavior, causing a feeding frenzy, some people will do dastardly things, and trolls will adopt the technology etc.As a response, virtually impotent lawmakers will bay at the moon.
ReplyDeleteI also suspect there will be an identity divide and some Luddite push back to the technology, especially since it, along with many other robotic levers will displace the roles humans have played since the dawn of time. In times of uncertainty and disruption, we have a tendency to get reptilian, dispensing with thought and acting on more instinctive levels.
I think we would do well to pave the runway before this plane fully lands, meaning contend with how to establish a meaningful existence once the necessity to work is largely eliminated. Social nature abhors a vacuum.
I could be missing something(s)
Joe Carter When the automobile was first invented, roads were just awful, so the automobiles were engineered with carriage suspensions and axles and wheels. Transmissions were geared for maximum traction, not for high speed.
ReplyDeleteThat wouldn't change until the building of the Lincoln Highway. Everyone understood the benefits of better roads, everyone except that old fuddy duddy Henry Ford, who thought the government should do everything.
Well, eventually the government did get in on the act, but not before the common-sense proposition of how improving the roads would benefit commerce of every description, a notion which went back to the Roman and Chinese empires, became firmly embedded in the culture at large.
The ultimate exponent of improved roads was Ferdinand Porsche, who engineered the Autobahn system, which would give rise to America's Interstate system. The cars had always been capable of running at 100+ km/hr. The math for the transmissions was as old as Archimedes. But the mass production of high speed automobiles created the need for the better roads, not the other way round.
Richard G Haven't you tried this ?
ReplyDeletehttp://www.adweek.com/creativity/man-creates-obnoxious-robot-trick-telemarketers-brilliant-quest-revenge-169373/
Dan Weese I would not see any reason to dispute any of that. Perhaps I made my point poorly, perhaps I misunderstand yours. If so, I apologize.
ReplyDeleteI was trying to speak to the potential social reaction to the confluence of technology on our collective psyche, not specifics of the trajectory of technology's march forward. When the industrial and thought revolution of the 18th, 19th and 20th century came to bear, it arguably gave rise to, or contributed to, some pretty self destructive events as a result of the various disruptions. Diasporas, displacements of all kinds and so on led to things like the Taiping Rebellion rebellion, and some of the disastrous wars of the 20th century which might not have been possible on those scales if people were not desperately looking for some sort of hope - some promise - some thing to hold on to that they could believe in for identity, safety etc. Powerful biological drives on which our basic social equilibrium exists. When we meddle in that space, we tend to get expressions of instability.
The point may just be far deeper than this format can carry. Or perhaps I am too inarticulate to convey it with sufficient compression without too great a noise loss, but I will give it a try,
We are on the threshold of perhaps the greatest disruptive social paradigm shift in not just recorded history, but evolutionary history. We will either contend with this head on, or probably get caught in the whirlwind.
Again, I could be missing something(s)
Rhys Taylor that is a thing of beauty. Where it cycles through excuses and says it needs the caller to repeat over and over. "So, local or national? Local or national?"
ReplyDeleteLovely.
Joe Carter ... nothing to apologise for, as I see it. I was just riffing on your excellent point about how we'll be forced to adapt to a world where the skinny kid who works in the kitchen won't be tasked with coming out front to talk to some customer in Hokkien because the sushi maker only speaks Cantonese.
ReplyDeleteRhys Taylor ... the telemarketers will adapt, too. It won't be humans calling:
ReplyDeletequora.com - Why does the porridge bird lay its eggs in the air?
Dan Weese My favourite bit of spam, context irrelevant :
ReplyDeleteOne of the zippered bags should be big enough to accommodate at least one Rainbow loom. The talk show hosted debuted the first (and probably only) full suit that included a jacket, tie and pants made of tiny rubber bands. You can learn something new yourself or give a gift of rainbows to a hobbyist you know.
In a former life, I wrote a modest system to assess and analyze a set of network abuse, and generate email reports.
ReplyDeleteSome of those went to the (presumed) responsible parties, or upstreams, some to other organisations tracking abuse, some to a public archive of such postings. This all ran automatically, with relatively little intercession on my part. I was sending so much email (reporting abuse), I'd joke, that I couldn't read most of it.
Those emails did clearly indicate that they were automatically generated. And aa not inconsiderable part of the system's complexity was in determining who not to send these messages to.
Found myself at a confab of (mostly) similarly minded folks. Heard from one person, representing a well-known and respected otganisation, but not identifying themselves personally (with reason), that my reports were often the difference between an action taken, or not, as mine provided independent verification.
And from someone else, working for a larger tech organisation, "I've written programs to respond to the emails your programs write." An interesting form of high praise from one geek to another.
If you're going to create an automated system, follow ssome common-sense, respect-based guidelines, and seek regulation requiring the same of others:
Clearly and unambiguously indicate that this is an automateed system.
Allow people to opt out. I'd recommend a parallel to national Do Not Call registeries for this purpose.
Require human-response call-back numbers.
Require registration and indication of such systems over voice comms networks.
Among the biggest losers of such systems, if unrestricted and widely adopted and inevitably abused are the relics of traditional voice comms sytems, both POTS (plain old telephone service) and wireless services. Increasingly, people are finding the negatives of both types of services, through telemarketing, robocalls, debt collectors, and mere wrong numbers, etc., to be so overwhelming that they're opting out of them.
(An underappreciated aspect of new telecoms technologies is that it's often their social exclusivity itself which is a major draw. Use allows and limits access to a refined club. With ubiquity, the frictionless universal contact becomes a net negative.)
Expect the extant telco monopolies to take a keen interest in this technology, likely from a regulatory angle.
An update. Google says that bots will identify themselves as bots. A university professor is horrified.
ReplyDeletehttp://www.bbc.com/news/technology-44081393
Prof Tufekci said she was surprised that the WaveNet project got as far as a public demonstration and wondered why it had not been quashed internally during development.
ReplyDeletea) I wonder if Prof Tufekci actually said that and/or what they meant
b) the answer to that question is transparency, public debate and the merits of research: you shouldn't refuse to develop a technology with unknown applications just because using it is unsettling and there are really no ethical concerns with being able to do the thing. Those come when you start using it in public.
Richard G She did:
ReplyDeletehttps://twitter.com/zeynep/status/994233568359575552
twitter.com - zeynep tufekci on Twitter
There is perhaps a perverse majesty to the value technological developments bring to our common table in that they have the power to leverage great benefit or destruction depending on how they are applied. This fact forces us to contend in ever greater clarity and impending consequence with the fact that without a commitment to each other penetrating to the deepest levels of society, we will be increasingly vulnerable to a mortal threat imposed by a momentary myopic lapse. While this might be terrifying, it also forces us to contend with the fact that we need to cultivate that shared interdependence on each other's commitment to each other as a matter of life and progress, or we could perhaps all suffer in the wake of that neglect.
ReplyDeleteThis may be nature's mechanism of apoptosis for sentient species who do not recognize the common interests that must be served to progress past the agrarian and industrial ages to a social one.
I could be missing something(s)
Edward Morbius “horrifying” is one thing, “shouldn’t have been invented” is another. Like demo and rollout. Admittedly, the gap between those can quietly disappear...
ReplyDelete