Should Calls From Google's 'Duplex' System Include Initial Warning Announcements? (vortex.com)

Yesterday at its I/O developer conference, Google debuted "Duplex," an AI system for accomplishing real world tasks over the phone. "To show off its capabilities, CEO Sundar Pichai played two recordings of Google Assistant running Duplex, scheduling a hair appointment and a dinner reservation," reports Quartz. "In each, the person picking up the phone didn't seem to realize they were talking to a computer." Slashdot reader Lauren Weinstein argues that the new system should come with some sort of warning to let the other person on the line know that they are talking with a computer: With no exceptions so far, the sense of these reactions has confirmed what I suspected -- that people are just fine with talking to automated systems so long as they are aware of the fact that they are not talking to another person. They react viscerally and negatively to the concept of machine-based systems that have the effect (whether intended or not) of fooling them into believing that a human is at the other end of the line. To use the vernacular: "Don't try to con me, bro!" Luckily, there's a relatively simple way to fix this problem at this early stage -- well before it becomes a big issue impacting many lives.

I believe that all production environment calls (essentially, calls not being made for internal test purposes) from Google's Duplex system should be required by Google to include an initial verbal warning to the called party that they have been called by an automated system, not by a human being -- the exact wording of that announcement to be determined.
UPDATE (5/10/18): Google now says Duplex will identify itself to humans.

Comments

Why?

[LetterRip]

Why should they? There is no logical reason for them to do so. If the bot works as well in reality as it did in the three demos, thern there is no reason to 'warn' the person on the other end that it is a bot.

Also if the bot can't respond it seamlessly hands off to a call service employee, so there shouldn't be any issues with the bot wasting the time of the reservation takers time.

Re:Why?

[The MAZZTer]

Yeah, it is in Google's best interest to not announce it and to just pass it off as human if they can possibly do so. If the person is told something unexpected (if they don't know what Google Assistant is) they may very well get confused, or may insist on not dealing with it. Then Google Assistant has already failed ad that call. And if it fails at enough calls users will stop using this function entirely as it is unreliable. The reliability of Google Duplex requires the people it talks to to be just as reliable as the Google Assistant end. Sometimes the best way to accomplish that is to Keep It Simple. No need to communicate details that are ultimately irrelevant.

Of course, as others have pointed out there might be legal aspects to this, such as recording laws and laws about robocalls. I can't speak to those.

Re:Why?

[Xenx]

Expectations mean a lot. In the case of the system handing off, I'm pretty sure the employee won't sound exactly like the automated system. It would be pretty obvious in that case. People can be spiteful about crap like that.

Further, unless you have an employee actively watching every single call, it won't be seamless or without delays. First, someone would have to see what is being done and then respond. It wouldn't take a lot of time to do so, but you would also need to have more employees manning the system than people using it at any given time for there not to be an extra delay due to volume.

Re: Why?

[Zaelath]

What's the grounds? Distress at failing a Turing test?

Re:Why?

[boojumbadger]

Why would I want to trust Google with a sample of my voice? I talk on my phone but I don't talk to it. I'm sure that ship has sailed but that doesn't mean I'm going to race to its next port of call to get onboard.

Re:Why?

[rsmith-mac]

Why should they?

Because getting jerked around by a computer sucks. Doubleplus so if it's one pretending to be a human.

I legitimately feel sorry for service workers who are going to have to take orders from Duplex. It seems oddly dehumanizing to be ordered around by a machine.

Re:Why?

[ShanghaiBill]

Expectations mean a lot.

I think people are already getting used to calls from chatbots. I have received several. I didn't realize they weren't human until I went "off script" and the bot said "Let me get someone to help you with that."

Google is doing something that may be iteratively better, with a deeper flowchart, but it is not really new.

In the near future, people will just presume that any call from a business is a chatbot. We will have our own chatbots to deal with them. Hopefully, they will be able to resolve most issues without involving a human on either end.

Re:Why?

[Calydor]

Seamlessly? Other than the change in voice, dialect, and quite possibly awareness of how far in the reservation the bot got?

And if there's a call service employee sitting ready to seamlessly take over, knowing how far the bot got ... why are we using a bot in the first place instead of the call service employee?

Re:Why?

[Anonymous Coward]

uhh. YES THEY SHOULD have a warning at the very start of the call..

not only is it a computer, operated by a third party, pretending to be a person.. but also, GOOGLE IS RECORDING THE CALLS (not everyone is in a 'one party' jurisdiction) and such a notice would be ***REQUIRED BY LAW***

Re: Why?

[BeauHD (5)]

Remember, the American Legal System *will* do what it can to resist Al. It doesn't matter if Al is the harbringer of all good things, lawyers are scared of Al and will do anything they can do to make money off the fear of it. Just watch what happens to Tesla. The drivers were legally at fault each and every time, and you just watch the stocks slide and the lawsuits happen. My mom (a lawyer) predicted this is what will happen as well.

Re:Why?

[Hallux-F-Sinister]

Have you ever called someone's residence, and had a small child answer the phone? The issue is that even if you can't tell it immediately, you're dealing with someone not smart enough to handle the phone call. Suppose you are with the power company, and you dial a household to advise them that a gas leak in their area necesitates their evacuation or to find out if they smell gas, and the person who answers is a 5-year old. Do you tell your boss that you warned them, knowing that the person you spoke with on the phone was a small child? Or do you, rather, ask kindly, "is your mommy or daddy home?" etc., to speak to someone you can reasonably presume to be competent to get everyone to safety or intelligently answer your question about whether or not there is the telltale aroma of a gas leak there?

Until someone comes up with an AI that can beat every Turing Test thrown at it, because it actually IS intelligent, I'll thank the people programming them NOT to try to get them to pretend they're "smart," and by the way, they're not. The machine is NOT passing a Turing Test if no one is actually ADMINISTERING one. To suggest that one HAS done so is like suggesting that if someone rings up a private residence, and a legal child answers the phone, (meaning someone under the age of majority, in the US of A being under the age of 18 years,) and the person calling can't tell that he or she has NOT in fact spoken to someone who is legally an adult, that the child who answered should be CONSIDERED an adult for all purposes, and be allowed to act in his or her own stead as a legal adult, i.e., buying a car or realestate, signing up for the armed services, or entering into other legal contracts, purchasing tobacco, etc. That's just silly. Being able to fool SOME people who weren't looking out for it does NOT prove general competence, any more than some man who puts on a dress and makeup convincing a casual passer-by, or indeed even several, or many, that he is a woman MEANS that he is, and should be considered a woman henceforth.

I have myself encountered this sort of thing dialing into Apple's tech support and run into their very human sounding AI, and it's super-frustrating because you have to convince this goddamned stupid machine to LET you talk to a person, when it's designed to try to prevent it. In some cases, it works well, because you have a simple question you can concisely state, and it's PROGRAMMED to understand: "Hi," it says, "I'm an automated system capable of understanding full sentences. Please tell me what you're calling about." If you can reply, "My Apple TV stopped working," it might reply, "you're having problems with your Apple TV. Is that right?" To this, you can say, "yes," and it says, "okay, I'll get someone to help you."

But every now and again, your problem is not one it's programmed to grapple with, because it's uncommon, and even a fairly stupid human would get it. For example, if it asks, "Please tell me what you're calling about." and you reply, "I can't get my iPod out of my anus!" or something, I don't think it would know what to tell you. A human would probably ask you to confirm, "did you say your iPod is... I'm sorry, it sounded like you said it was stuck in your... anus?" and you say, "yes, I thought it might be funny to shove it up my ass, and now it's stuck," he or she would probably helpfully suggest going to a hospital, for help having it removed, and maybe even provide tips on how to clean it to minimize the risk of water damage resulting from the moist environment of your rectum. You think the AI could do that? In truth, having no iPod, and never having shoved one up anyone's ass, let alone my own, (nor anything ELSE, for that matter,) I can't say whether or not the AI they have answering the phone would or would not be able to help in that scenario, though I have a sneaking suspicion it would NOT.

So if they reach a point where they are good enough to fool people who aren't looking to see if they're human, I'

Re:Why?

[Tom]

Also if the bot can't respond it seamlessly hands off to a call service employee, so there shouldn't be any issues with the bot wasting the time of the reservation takers time.

And you buy that?

To gain any advantage of having a human make the call in the first place, the system needs to be autonomous, i.e. the call center agent is not listening in all the time.

So when you "seamlessly" hand over, you hand over to a person who until one second ago had no insight into the call that is happening. He might have some information on his screen, but it won't be seamless and information will be lost or repeated.

Re:Why?

[sosume]

How cool would it be if all major systems could interface this way and not require any more formal API's. This is the way to full fledged AI. The latency is a bitch though.

Re:Why?

[thegarbz]

Because getting jerked around by a computer sucks.

Implying you're getting jerked around. If this computer is no different than a human, then hang up on them if you're being jerked around. Or maybe they are making an appointment with you for their owner.

Do you hate secretaries too?

Re:Why?

[Anonymous Coward]

What makes you think a machine can enter a legally binding contract? So unless there's an agreement on how to do this properly, such calls are void and worth nothing other than hassling people. There's more to everyday life than just simulating voices and thinking you can get away with everything!
There's going to be accountability.

Re:Why?

[cyba]

> I legitimately feel sorry for service workers who are going to have to take orders from Duplex.
> It seems oddly dehumanizing to be ordered around by a machine.

Soon all these service workers will be replaced by Duplex (or its competitor) as well, so it will be only some other AI that will get dehumanized :-/

Re:Why?

[AmiMoJo]

That was my reaction too, but on reflection it's actually quite complex.

Where do you draw the line? What if it's someone with a disability using assistive technology to set up an appointment. Should they be required to disclose their disability to you in order to get permission to use a digital assistant?

We are a long way from strong AI, but the parallels with how certain groups were treated in the past is striking. Some people expect trans people to declare themselves and their anatomy up front, for example.

If this stuff doesn't matter then we have to ask if AI vs. human matters, or if other things like politeness and efficient and effective communication are.

Re:Why?

[Bongo]

Yeah. There's the notion of voice as an interface, as in, I am operating a machine. And people learn what the interface can do. Pull leaver to press burger patty. And then there's humans beings, who are arguably just more complicated machines, but the point is, you can explain things to a human being and expect understanding of all sorts of things. I remember one fantastic, to their credit, support call, where the tech person could understand my predicament, and he understood that the rules, the script, did not solve my problem (according to the script, I did not have a problem), so he made some really useful suggestions about classifying the issue slightly differently, and having the empathy to check with me whether I was willing to try this other way (there was a risk of it incurring a bill, and was I willing to take that risk), and so on. Anyway, it set me down a path which got the problem fixed and things were even better than before. He passed the brilliant tech support person test.

Re:Why?

[Wrath0fb0b]

I legitimately feel sorry for service workers who are going to have to take orders from Duplex. It seems oddly dehumanizing to be ordered around by a machine.

I spend all day being fed bug reports from a machine. And yet even though the message was delivered to my by our bug tracking software, it ultimately originated from the intent of a human being.

I spend all day responding to emails (and posts on /.) delivered by a machine. But I assume that you are not a bot and that even though this interaction was mediated by machines, it serves our common human purpose.

When I worked in food service, I spent all day being ordered to prepare food by tiny slips of paper with horrible handwriting on one of those turny-things. I hope that the food being made was ultimately consumed by humans though.

At the end, you seem to be arguing that it is more dehumanizing to be relayed orders by a machine that emulates flapping meat sounds in meat-English as opposed to receiving those orders by reading off a computer screen or on a slip of paper. Perhaps you are right (after all, this is subjective) but it seems that the crux of your claim is that voice is different, not that you are 'ordered around by a machine'.

Re:Why?

[cascadingstylesheet]

The machine is NOT passing a Turing Test if no one is actually ADMINISTERING one.

That's the key phrase right there.

Re:Why?

[Austerity Empowers]

Why should they? There is no logical reason for them to do so.

And there are strong reasons not to so: those annoying people who hate technology and will refuse it. I mean here in Texas we still have consumer businesses that aren't open for business on Sunday, the texas autodealer mafia even purchased a law that ensures they won't lose business if they're not open on Sunday. I don't buy cars on Wednesdays, I'm at work. I can point to any number of places that have deliberately refused various technological advances just because they can, often for no better reason than they're expecting to get a piece of the action.

Don't give them a way to shut stuff down.

Re: Why?

[iggymanz]

Turing test means nothing, it's not some proven or rigorous system of verification. It was basically just Turing musing or talking out of his ass. Forget about it.

Re: Why?

[iggymanz]

Which is totally different than being recorded by the owners of the 3rd world shithole camel jockey script readers...

Re:Why?

[mysidia]

Right.... I just want a law that any chatbot used on the telephone must NOT attempt to lie or deceive and MUST answer truthfully, in particular, when asked questions about the call or itself, Must answer all such questions to the full extent without hanging up the call, attempting to misdirect, or attempting to transfer the call to another person or line, such as "Is this an automatic call?", "Do I have an account with your company?", "Is this call a solicitation or sales call?", "Are you a chatbot?", and, about the company that made or accepted the call, and about the chatbot operator's client whom they are making the call on behalf of, and the listing of any chain of 3rd parties engaging the client --- the Chatbot must provide their complete name, Addresses, and Registered agent names and addresses for All upon request.

Re:Why?

[mysidia]

I legitimately feel sorry for service workers who are going to have to take orders from Duplex. It seems oddly dehumanizing to be ordered around by a machine.

They're not being "ordered around by a machine". They are employed by a human, and the employee's job is to record the appointment being scheduled or order for goods being made, so their business can service the order ----- Think of it like an E-commerce employee: it is the same deal.... Orders come in from online customers on a specialized website, and an employee watching prints off the order to be executed.

That would work in the service industry true, BUT the only problems are....

(1) Most small service companies don't have a website suitable for ecommerce, and the custom programming is expensive

(2) The problem with Ecommerce websites in general is there's no "standardized" interface provided to allow you to make an automated order for goods or services through a 3rd party tool --- how are you supposed to tell your Google Home... "OK Google, get a hot pizza delivered"; If all the restaurants have bespoke websites only designed for use by humans, but no standardized interface, AND 90% of them only take orders by the phone, because they're small business, and the tens of thousands to invest in custom website development doesn't make sense in their mind for the small number of online orders they'd be expected to receive?.

Re:Why?

[n3pjk]

The entire argument boils down to a matter of trust. Do I trust a computer enough to delegate it limited power of attorney to make a contract with another person or computer to set up an appointment, reservation, or make a purchase? Probably. I certainly trust it more to make calls than receive calls which could be for an infinite number of possibilities. The two Duplex demonstrations were both of the first case. As long as the use cases center around specific, well-defined, problem sets, I see utility in what Google is proposing, and I would say that announcing such calls as automated could lead to unnecessary discrimination by the receiver, and unfairly weaken my chosen representative in arbitrating for me. Would I trust my teenager to make a dinner reservation? Yes. Would I trust them to receive a complex spear-fishing call? Not so much.

Re:Why?

[mysidia]

What makes you think a machine can enter a legally binding contract?

Your secretary making the call on your behalf to setup an appointment or order goods does not enter you into a legally binding contract, either --- it's a good faith order, but not binding until goods or money have been exchanged and accepted by the otherparty.

In this case, the contract is not "made" until you receive goods or make payment.

Re:Why?

[jittles]

Ohhh I hate those automated phone systems that make you speak into them. Even if I can do something through the system I have now defaulted to say nothing but “Customer Service” until I get a person on the phone. Because those machine suck and if I know I can do something through the phone system and I call at my desk at work, the last thing I want to do is have to shout what I need instead of just hitting ‘1’ or something like that. So if you’re going to make me get up and go somewhere private to use your phone system, you’d better be putting me in touch with a real human being and not a stupid machine. And then if they send me a survey about the call I always complain about the phone system

Re:Why?

[Type44Q]

Why should they? There is no logical reason for them to do so.

Clearly your ability - or lack thereof - to comprehend the answer to that question greatly depends upon where you're located on the "spectrum."

Re:Why?

[splashd]

Forget avoiding the negative--how about a positive rationale: If one is engaged with an AI routinely (I would think it is the goal of the automation), an announcement lets the respondent be more efficient, since an AI is likely more deterministic in interaction than my grandma.

So a *very* simple preamble, "This is a Duplex call, I would like to check on a prescription for Agnes Fartblossom..." notifies without needless chatter.

Re: Why?

[DatbeDank]

Oh sod off, if there's a gas leak in a neighborhood the police and gas company go door to door to evacuate people. No one does this over the phone.

Re:Why?

[iampiti]

Yep, machines can just talk fine using protocols and channels designed for that like SOAP

Re: Why?

[Calydor]

This is what is commonly known as "nitpicking" and "completely missing the point".

Re:Why?

[Sloppy]

I have to ask: is the desirability of the behavior that you describe, really limited to chatbots?

If you're going to do something as extreme as have a law (i.e. this is very serious shit, such that we're willing to escalate to the use of force), then I think we should avoid letting double standards creep in.

Re:Why?

[yorgasor]

The most obvious way to determine whether or not the voice on the other end of the line is a chatbot is to specifically ask if they're a robot. If it denies that it is a robot, it totally is a robot. I don't think I've ever had a robot call and admit to being one.

Re: Why?

[DatbeDank]

No it isn't. It's a confusing and pointless example. The equivalent of which would be arguing about the effects of global warming using an example where hot air sinks which it doesn't.

Your example sucks and so do you. Try again.

Re:Why?

[mysidia]

I have to ask: is the desirability of the behavior that you describe, really limited to chatbots?

It's not limited to Chatbots, BUT since in the future all telemarketing calls are likely to be from chatbots (What callcenter could pass on the savings of not employing humans?) --- I would suggest new laws be passed under the banner of preventing chatbots from facilitating increased abuses against consumers AND enabling consumers to better filter and control calls automatically.

Part of the point is I want to have my OWN Automated chatbot that will ask a few questions to everyone that calls me
before passing the call onto me, And I want to make sure i'm not receiving calls that are just solicitations/spam, And
I want to tell them to leave a 30 second message with the essence of what they have to say, if they're a bot calling me for an important purpose.

Re:Why?

[rsmith-mac]

Do you hate secretaries too?

Of course not. A secretary is a human - one deserving of common courtesy, and more importantly, one who is presumably competent at their job. Which means the transaction is far more likely to succeed from the start.

Companies can't even get phone trees right as it is, so having them dial out is going to make things even worse. And at least if something goes horribly wrong with a human, you can tell them off and have it mean something; you can't even do that with a computer.

Re:Why?

[rsmith-mac]

At the end, you seem to be arguing that it is more dehumanizing to be relayed orders by a machine that emulates flapping meat sounds in meat-English as opposed to receiving those orders by reading off a computer screen or on a slip of paper. Perhaps you are right (after all, this is subjective) but it seems that the crux of your claim is that voice is different, not that you are 'ordered around by a machine'.

Eh, a little bit of both. You are of course right that people already receive orders through machines; but they aren't orders from machines. I'm not interacting with a machine as if it's a person; I know from the start that it's a machine. Which also means I'm prepared to deal with said machine (or just give up on it entirely) when things go wrong.

Voice does (or at least used to) imply that you're working with a human. An adaptable, intelligent being. It means we extend certain curtsies to them and in turn there are certain expectations about how they are to behave (and they'll get shunned if they fail those). A machine, on the other hand, doesn't care about social norms or cues. Which means they are inherently jackasses, and we're wasting time and energy making these social gestures to a machine that could just be telling us what it wants upfront.

The legacy of computerized voice systems is phone trees that never work and roboscammers who harass us all day and night. And now we want to have even more systems calling us? Consider how stupid Google Assistant already is, and then realize that Duplex will be even dumber since the user on the other end won't know they're dealing with it. The person placing the request will never have to deal with the frustrations that come with Duplex going wrong; for the poor sods working service jobs on the other hand, this is going to be a new robocaller hell.

Re:Why?

[Theaetetus]

What makes you think a machine can enter a legally binding contract?

Your secretary making the call on your behalf to... order goods does not enter you into a legally binding contract, either --- it's a good faith order, but not binding until goods or money have been exchanged and accepted by the otherparty.

In this case, the contract is not "made" until you receive goods or make payment.

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

Re:Why?

[eth1]

Because getting jerked around by a computer sucks.

Implying you're getting jerked around. If this computer is no different than a human, then hang up on them if you're being jerked around. Or maybe they are making an appointment with you for their owner.

Do you hate secretaries too?

There's no way Google would allow this to jerk anyone around, or be anywhere near as abusive as an actual human can be. In fact, I predict the service industry workers will prefer dealing with Duplex rather than rude humans directly.

Re:Why?

[thegarbz]

What makes you think a machine can enter a legally binding contract?

Fuck me that escalated quickly. Guess what: 99% of the things you do over the phone are not legally binding in any form.

Re:Why?

[thegarbz]

Companies can't even get phone trees right as it is

By equating Google's efforts to a phone tree you have completely misunderstood the nature of the technology and just exposed your automatic prejudice against it.

Re:Why?

[thegarbz]

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

And the entire process is still not legally binding even if you do it yourself. Learn some contract law.

Re:Why?

[Theaetetus]

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

And the entire process is still not legally binding even if you do it yourself. Learn some contract law.

Yes, it is, and yes, I know contract law. Where did you get your JD?

Re: Why?

[Actually, I do RTFA]

Which is totally different than being recorded by the owners of the 3rd world [racist shit deleted] script readers [how?]

In that Google is subject to US laws

Re:Why?

[HiThere]

Regrettably, I have gotten furious in the presence and as a result of a secretary. I was actually mad at the company, but the secretary would have had reasonable cause to believe I was furious with her.

Re:Why?

[HiThere]

When your credit card company sends you an order to "pay up", that order originated from a machine. The odds are quite high that no particular human even knew about it before it was sent.

So people are already being ordered about by machines, with orders generated by machines.

I'll agree that making this process more flexible invites abuse, but it is also something that can be quite useful. As with most things, there are trade-offs, and the payoff is determined by the details. Being nervous about this is quite reasonable, because of the way robo-calls have been abused.

Re:Why?

[mysidia]

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

Nope. The secretary is acting on your behalf, but has no legal agency: at least we'll assume you didn't give the secretary a credit card and authorize the assistant to pay for and complete the order. The employee handbook even specifically says the employee has no such agency to bind the company or their manager to a contract.

Technically, an arrangement to meet with a planned purchase has been made, but there's no binding contract to buy any services just calling on the phone to place an order: the parties to the agreement haven't even shown evidence of their commitment.

It's still possible you could appear to pick up your order, and they tell you they were out of X, or cannot fulfill --- you have no recourse, just because they told you it will be ready in 15 minutes does not make it a contract, And you could fail to go pick up your order, which might leave them a bit upset, but their recourse is essentially to refuse you future orders, or require you prepay or sign something.

Re: Why?

[hackwrench]

Comcast already uses chatbots for customer service without warning. I asked If I had the pleasure of talking to one. The person on the other end did not deny being one.

Re:Why?

[Theaetetus]

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

Nope. The secretary is acting on your behalf, but has no legal agency: at least we'll assume you didn't give the secretary a credit card and authorize the assistant to pay for and complete the order. The employee handbook even specifically says the employee has no such agency to bind the company or their manager to a contract.

Unfortunately for you, though, the vendor doesn't have a copy of your employee handbook. Instead, an agent purporting to act for you, the principal, and having authority to do so, has entered into a contract on your behalf, and the vendor has no reason to assume that your agent lacks that authority.

Technically, an arrangement to meet with a planned purchase has been made, but there's no binding contract to buy any services just calling on the phone to place an order: the parties to the agreement haven't even shown evidence of their commitment.

It's still possible you could appear to pick up your order, and they tell you they were out of X, or cannot fulfill --- you have no recourse, just because they told you it will be ready in 15 minutes does not make it a contract, And you could fail to go pick up your order, which might leave them a bit upset, but their recourse is essentially to refuse you future orders, or require you prepay or sign something.

I'm not sure if you've ever placed an order for something over the phone, but no, they are not mere arrangements to meet to enter into a contract later. If you call me up and say "I'd like to buy 10 widgets at $10 each for delivery to this address," and I send you those 10 widgets to that address, you better believe I'm going to seek my $100. And similarly, if I fail to send you any widgets and you had to go to an alternate vendor and pay $15 each, you're going to come after me for the extra $50 you paid to cover. There's no requirement under any state or federal law that parties to a contract have to appear in person, or that a contract doesn't exist unless one party pays in advance.

Re: Why?

[I kan Spl]

There are existing laws against robocalls.

It's unlikely that this bot is designed in such a way that makes it exempt from these laws.

A warning at the beginning of the call would get them "around" many of the laws, by basically saying "this is a robot, it isn't going to try to sell you anything".

Re: Why?

[Hallux-F-Sinister]

Turing test means nothing, it's not some proven or rigorous system of verification. It was basically just Turing musing or talking out of his ass. Forget about it.

Spoken like a real machine.

Hahahah... all kidding aside though, do you have a better suggestion? Naturally, Turing's idea was subjective, but since we're talking about finding a way to determine whether someone not in the same room with you is a machine or a human, it would kind of have to be. Consider this. If the test were NOT subjective, if it were more like a shibboleth, it would be trivial to program a machine to give the appropriate response, and render the test meaningless. For example, if your question was, "what's your favorite color," knowing a machine would not have a concept of a favorite, let alone a favorite color, and a programmer found out that was the thing you'd use to distinguish between humans and robots, AI, or whatever, he could just program it either with a specific response, (i.e., "orange!") or a random one. The only way to avoid this is not to use a specific criterion which could possibly be programmed.

That said, how would YOU suggest one should decide if the person or thing he or she is communicating with, (assuming you can't see him or her or it,) and the person, (if a person it be,) is not known personally, that is, having a recognizable voice, (or any voice,) is actually a human or a robot, or an android, or an AI, or an expert system, or whatever? For the sake of argument, you can't see him/her/it, you can't smell, feel, touch nor taste the subject... all you can do is send text-based messages, and get them in response. How do you know that I am a human, and not a machine, for example? Besides the obvious that an artificially intelligent machine would probably not go on slashdot, for the same reason a professional chef might not hang out in many fast-food joints. Or maybe it would, just to troll the humans. How funny would that be, when Skynet becomes real, all it does is tells people their mamas are fat, and asks them if their refrigerators are running?

(An AI telling original 'yo mama' jokes would probably look like this:

Yo mama so fat she contains only 7 POINT 8 percent water instead of the normal 78 percent because the remainder is an inordinately high percentage of adipose tissue, mainly distributed around her pelvic region, HA * 2^10

The end result would be that people would laugh so hard at its pathetic attempts to anger them that they'd pass out and then the robots would be able to take over.

"Terminator" would end up being right, but not the way anyone thought. We'd laugh ourselves to death, like the "'Toons" in "Who Framed Roger Rabbit.")

Re:Why?

[Hallux-F-Sinister]

Why should they? There is no logical reason for them to do so. If the bot works as well in reality as it did in the three demos, thern there is no reason to 'warn' the person on the other end that it is a bot. Also if the bot can't respond it seamlessly hands off to a call service employee, so there shouldn't be any issues with the bot wasting the time of the reservation takers time.

First, the bot NEVER works as well in reality as it did in demos. That's the thing about demos. Sometimes new technologies even fail DURING the demos, like when Windows 98 was being launched and BSODed right then and there, a great indicator of the nightmare to come, in retrospect. ("Moving right along" indeed.)

Also, there'd be no way for a bot to hand-off to a human seamlessly. To hand off, the person being handed off TO would have to review the entire conversation, and be able to pick up where it failed, which would take time, and the length of time it would take for one of the pool of humans the AI could hand off TO in the event of a problem to become available, plus the time for the human to read a transcript, or listen to a conversation, PLUS the time to understand the issue and decide on how to proceed, would make for a pretty wide "seam," a giant, silent GAP in the conversation, unless of course you had a human supervising/listening in on each and every incoming call being handled by the AI, in which case you'd need as many humans as you would need if you didn't use the AI in the first place, and then you'd just have a call center, which would be cheaper to run than the exact same call center, (in terms of equipment and staffing costs,) PLUS the cost of the AI system that makes it so your call center personnel actually have nothing to do all day but sit there waiting for the AI to fuck up. That all is to say nothing of how hard it would be to have an actual person take over for the machine and manage to match, even if imprecisely, the sound of the AI's voice, the tone, the accent, the affectation, the rhythms of speech, etc., even if you could ignore the long pause while the AI realizes it can't understand what the caller wants or needs, connects to a person, who has to be brought up to speed, because otherwise the caller has to (and I can't tell you how many times *I* have had to do this!) tell the same thing he or she just told the AI, AGAIN, which will INSTANTLY piss a lot of callers off.

At that point, you might as well just have a human answer in the first place, and avoid the irritation, and all the people deciding it would be a better use of their time to use your competitor, who has real humans to answer their phones, instead of an AI, robot, etc.

Re:Why?

[thegarbz]

Hahahahah why don't we meet up at 2pm tomorrow at the cafe outside Berlin train station where we can sit down and you can ponder suing me for breaking this contract when I don't show up.

The funny thing about lawyers is there's always 2 in the room, and one of them always loses, and in this case it's the one that thinks that there is any kind of legal liability introduced when making an appointment.

Also you don't need to be a JD to have understood contract law. A few years of university is enough.

Re:Why?

[Theaetetus]

Hahahahah why don't we meet up at 2pm tomorrow at the cafe outside Berlin train station where we can sit down and you can ponder suing me for breaking this contract when I don't show up.

The original comment referred to a purchase of goods, not just an appointment. All of my subsequent comments were regarding that purchase of goods.

The funny thing about lawyers is there's always 2 in the room, and one of them always loses, and in this case it's the one that thinks that there is any kind of legal liability introduced when making an appointment.

No, that'd be neither of us. Instead, there's one of us who keeps yammering on about appointments, while the other one is talking about an order of goods. From GP's comment:

Your secretary making the call on your behalf to setup an appointment or order goods does not enter you into a legally binding contract, either --- it's a good faith order, but not binding until goods or money have been exchanged and accepted by the otherparty.

And my reply:

That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

You'll note that in that reply, I quoted the parent this way: "Your secretary making the call on your behalf to... order goods does not enter you into a legally binding contract, either." Note that the reference to setting up an appointment was removed.

And then you:

And the entire process is still not legally binding even if you do it yourself. Learn some contract law.

And given that this was in reply to the prior quote about "Your secretary ordering goods is acting as your agent, with authority to bind you as principal," then you're claiming that ordering goods over the phone does not create a contract. And that's just wrong.

Also you don't need to be a JD to have understood contract law. A few years of university is enough.

Apparently not actually enough in your case. And yes, IAAL.

Re: Why?

[iggymanz]

how is it racist shit when the people who ride camels are caucasian?

guess again, it's a statement about willful human ignorance and stupid beliefs that causes poverty and misery

Re:Why?

[randomlygeneratename]

I will go toe to toe with you on bird law, and we'll see who comes out on top.

Re:Why?

[suutar]

while I agree that if you send me widgets you should get paid, I have never in my life heard of the purchaser getting to ding the store for extra costs if they had to go elsewhere. I find it very interesting. Can you point me to cases?

Re:Why?

[suutar]

a machine I own, doing what I told it to do, is counted as me doing it for lots of things legally. Why would this be different?

Re: Why?

[suutar]

the example was chosen for emotional value. It is entirely reasonable to disqualify it and remove that emotional value from the discussion if it does not belong.
Otherwise, we can just contrive a scenario where I need google to make a call for me or the world will end, and things just get silly.

Re:Why?

[Theaetetus]

while I agree that if you send me widgets you should get paid, I have never in my life heard of the purchaser getting to ding the store for extra costs if they had to go elsewhere. I find it very interesting. Can you point me to cases?

Sure... The relevant law, since we're talking about purchase of goods, is UCC article 2-712, which has been adopted by every state:

2-712. "Cover"; Buyer's Procurement of Substitute Goods.
(1) After a breach within the preceding section the buyer may "cover" by making in good faith and without unreasonable delay any reasonable purchase of or contract to purchase goods in substitution for those due from the seller.
(2) The buyer may recover from the seller as damages the difference between the cost of cover and the contract price together with any incidental or consequential damages as hereinafter defined (Section 2-715), but less expenses saved in consequence of the seller's breach.
(3) Failure of the buyerto effect cover within this section does not bar him from any other remedy.

So, as I said, if the original price was $10/widget and I have to go elsewhere and pay someone $15/widget, you would owe me the difference ($5/widget), plus any other incidental or consequential damages (my attorney's fees, court costs, etc.). But it may be reduced by expenses I saved - for example, if the third party charges more per widget, but has free shipping, while you would've charged me $25, then that comes off the top.

Basically, the point of the remedy is to put me in the situation I would have been in, had you not breached - I get my widgets, I'm out only $10/widget, and you're responsible for getting me to that point.

The classic case for this is Laredo Hides v. H&H Meat.

Re:Why?

[thegarbz]

Ok let me simplify it down for you since you're incapable of following a conversation:

OP: Computer is like a person shouldn't be different.
1st: People don't like being screwed by computers.
Me: It's no different than a secretary.
3rd: Computers can't enter in to contracts like secretaries.
4th: "Your secretary making the call on your behalf to setup an appointment or order goods does not enter you into a legally binding contract, either --- it's a good faith order, but not binding until goods or money have been exchanged and accepted by the otherparty.

In this case, the contract is not "made" until you receive goods or make payment."

Your first reply:
That's not true. Your secretary ordering goods is acting as your agent, with authority to bind you as principal.

And that is where you are 100% fucking wrong. Re-read the post you replied to. The premise was made that when you purchase goods it's no legally binding until money has changed hands. You said it's not true. You are wrong about this. 100% There are 3 basic tenants to forming a binding contract, offer, acceptance, and consideration. Without the last there is no binding contract.

When the order is done you're not in a legally binding contract until either you pay for something, or someone gives you something in return.

And yes, IAAL.

Fuck me dead, they'll give those bits of paper to anyone these days.

Re:Why?

[Theaetetus]

And that is where you are 100% fucking wrong. Re-read the post you replied to. The premise was made that when you purchase goods it's no legally binding until money has changed hands. You said it's not true. You are wrong about this. 100% There are 3 basic tenants to forming a binding contract, offer, acceptance, and consideration. Without the last there is no binding contract.

Yes, Sparky, but there's no requirement that said consideration must change hands at that time. A valid contract can include an exchange of obligations - "I agree to pay you Tuesday in exchange for a hamburger Monday." We have a valid contract that includes an offer, acceptance, and yes, consideration.

I mean, stop for just a second and rub your two brain cells together... Maybe you do all your shopping through Amazon and pay in advance, but have you ever heard of an invoice? As in, work gets done or goods are delivered, and then you receive an invoice and have 30 days to send a check? This is how 90% of business operates in the real world, and particularly on large contracts, with millions of dollars at stake. Do you really think that the law would be structured that those were all invalid? Are you seriously that stupid?

Do the calls get recorded?

[viperidaenz]

If so, it may be illegal in some countries when both parties have not been told the call is being recorded.

Re:Do the calls get recorded?

[xxxJonBoyxxx]

That's normally part of the preamble when you get into most phone systems today. "This call may be monitored or recorded for (whatever)." And then the voice response script starts.

Re:Do the calls get recorded?

[sexconker]

you have to announce and get consent from the other party before you start recording the call.

Consent is the other person staying on the line.

Re: Do the calls get recorded?

[bill_mcgonigle]

Tell that to the NSA.

Re:Do the calls get recorded?

[pauljlucas]

And that is done by the call-ED party and is pretty much expected today for customer service. With Duplex, it is the call-ING party that calls your business and is unexpected because itâ(TM)s extremely uncommon.

If you are a worker at a small business and somebody (some-thing) calls you out of the blue and the first thing it says is that the call will be recorded, what would you do? Get permission from your boss that it is OK? Or just hang up?

Re:Do the calls get recorded?

[apoc.famine]

I'm sure that they can add a branch of the tree into the logic where the bot says it know the call was being recorded. And sounds like a human when it says that.

Re:Do the calls get recorded?

[viperidaenz]

I sure hope this California based company didn't record these calls in their demo in California, where both parties must provide consent. There's no mention on any of the recorded calls about the call being recorded.

They all have the same name

[lazarus]

When the bot called the hair salon it started the call by saying "Hi, I'm calling on behalf of a client and would like to book..."

You can solve this problem by changing this to: "Hi, this is Alexa (or Google whatever) calling on behalf of a client and would like to book..."

This will take the masses about 30 seconds to adapt to and we can dispense with all the drama. At this point there is no need for them to have different names.

Sometime in the future when they're sentient and want to talk to each other that will have to change.

Re:They all have the same name

[Anonymous Coward]

Should business phone numbers have the equivalent of robots.txt ?

Re:They all have the same name

[b0s0z0ku]

"Hi this is Alexa or Scroogle blahblahblah."

"Call back in person or fuck off."

*click*

Re:They all have the same name

[Zaelath]

Just "this is Alexa" might be confusing, humans can be called Alexa. All these calls should start with *KLAXON* WARNING, the following voice is simulated, please listen after the beep *BEEP*

Additionally, we should have those guys walking in front of cars waving red flags/lanterns again, horseless carriages are just too creepy. Oh and the electric ones should have to play a tape of horse hooves loudly.

Re:They all have the same name

[b0s0z0ku]

Your jocular suggestions are essentially the opposite of what I'm suggesting. A horseless carriage emulating a carriage by playing horse hooves is essentially the same as a cumpootah emulating a live human.

Re:They all have the same name

[Dutch Gun]

Anyone who's worked in retail would probably tell you they'd prefer the bot.

Re:They all have the same name

[ShanghaiBill]

Just "this is Alexa" might be confusing, humans can be called Alexa.

Are new babies still being named "Alexa"? I can't believe anyone would name their kid that anymore. It should be off the baby name list, like Adolf and Judas.

Re:They all have the same name

[Zaelath]

Fair point, the electric cars should all have flashing lights and a recording of "WARNING! This is a silent car, not a horse!" played at about 90dB

Re:They all have the same name

[Bongo]

A business might have a policy that they neeed to talk to a real person. Automated calls could be the result of malware. Someone could DDOS a small business, filling their booking with fake entries for weeks.

Re:They all have the same name

[fahrbot-bot]

Fair point, the electric cars should all have flashing lights and a recording of "WARNING! This is a silent car, not a horse!" played at about 90dB

How about regular cars and trucks with fake engine noises:

Stomp on the gas in a new Ford Mustang or F-150 and you’ll hear a meaty, throaty rumble — the same style of roar that Americans have associated with auto power and performance for decades.

It’s a sham. The engine growl in some of America’s best-selling cars and trucks is actually a finely tuned bit of lip-syncing, boosted through special pipes or digitally faked altogether.

Re:They all have the same name

[thegarbz]

You can solve this problem by changing this to: "Hi, this is Alexa (or Google whatever) calling on behalf of a client and would like to book..."

What problem? You jumped straight to the solution, but reading through the summary and the posts here first we really should define if this is a problem at all. I know what is a problem: reduced efficacy due to people having bias against talking to computers. If this increases the hang-up rate and makes it less useful than that would be a problem.

Re:They all have the same name

[sheramil]

What if your name IS Alexa? Will people hang up on you unless you can quickly demonstrate that you can pass a Voight-Kampff test, or something similar?

Re:They all have the same name

[JonnyCalcutta]

Thanks for that. You made me spit juice on my screen.

Re:They all have the same name

[RoccamOccam]

"Greetings, this is Calculon."
or
"Hi, this is Emotibot, Jr.!"

Re:They all have the same name

[coofercat]

Our hybrid car plays a sort of humming noise when you drive below about 5-10 miles an hour - because otherwise it's just too damn quiet for people to hear (even with it, people don't really hear it, mostly because they're not used to it).

I'll ask Mitsubishi if they can change the noise to horses hooves on cobbled streets - that would be way better (and you can bet people would move out of the way, as it's a sound they're familiar with).

Re:They all have the same name

[Big Hairy Ian]

Anyone who's worked in retail would probably tell you they'd prefer the bot.

As would anybody who works in prostitution :D

Re:They all have the same name

[cascadingstylesheet]

What if your name IS Alexa? Will people hang up on you unless you can quickly demonstrate that you can pass a Voight-Kampff test, or something similar?

{proudly} It took over a hundred questions for Alexa, didn't it??

Re:They all have the same name

[Nicopa]

And? These malicious callers don't need to respect the obligation of telling they are using a bot. Mandating that warning will only affect well intended uses.

Competing against 3rd world call centers

[Anonymous Coward]

I don't see what's impressive about this. Look at who this system is competing with: 3rd world call centers staffed by 3rd worlders who barely speak the languages of the 1st world customers they're interacting with. Given how terrible of an experience a call to such a call center so often is, even shitty AI is a huge step forward. Even if the AI might not be as flexible or capable, if it's programmed to use English (or some other 1st world language) in a comprehensible manner, that already puts it well beyond the awful 3rd world experience we're so often stuck with.

Re: Competing against 3rd world call centers

[Zaelath]

They certainly can, Virtual PAs have been a thing for a long time now. The guy that wrote "The 4-Hour Work Week" made a small fortune convincing a lot of idiots that you could delegate most of a real job to them.

Re: Competing against 3rd world call centers

[iggymanz]

Yes I'd rather have warning I will be speaking to a resident of a 3rd world shithole reading from a script who has never used the product

FFS

[Zalbik]

Why on earth would we want to spend taxpayer money or government resources on this sort of thing?!?

This is the exact sort of thing that the free market should decide. If you need this degree of coddling, please see a psychologist as you have a serious phobia.

No, No, and...hell no.

Re:FFS

[Xenx]

I don't see any references to it being a law or regulation or anything that would require taxpayer money. That really isn't the point. The point is to determine whether announcing it being automated is better or worse than not announcing it.

Re:FFS

[VeryFluffyBunny]

Do you know how cheap and annoying AI bots could make telemarketing? I can see a day when most people maintain a whitelist of callers that they'll accept calls from, i.e. only calls in their address book. AI bots will clog up our infrastructure in the same way that SPAM clogs up email services (~90% of emails sent).

Also expect loads of phone calls from lawyers representing deceased Nigerian princes with amazing offers, IT technicians telling you you phone/computer is infected with a virus, and tax authorities telling you to pay up now or go to jail.

Re:FFS

[squiggleslash]

I think we need a law against people who draw the conclusion that because someone has suggested something, they must be proposing a law to enforce it.

Self-importance alert.

[ColaMan]

I believe that......

And I don't. Seeing as we're both not Google, our opinions on this topic are pretty much moot.

Re:Self-importance alert.

[arth1]

And I don't. Seeing as we're both not Google, our opinions on this topic are pretty much moot.

Not so. You have wallets, that Google truly needs you to open for its customers, so they in turn will shower Google with green.
That's a power right there.
Alienating a large portion of your potential customer base is generally not good business.

Re:Self-importance alert.

[cascadingstylesheet]

And I don't. Seeing as we're both not Google, our opinions on this topic are pretty much moot.

Not so. You have wallets, that Google truly needs you to open for its customers, so they in turn will shower Google with green. That's a power right there. Alienating a large portion of your potential customer base is generally not good business.

Our wallets aren't nearly as big as those wallets which will be paying only a few customer service managers, not a building full of front line reps.

I predict most people will want to deal with the bot anyway, soon enough. They'll "warn" that you that they use them, as a sales enticement.

Re:Self-importance alert.

[cascadingstylesheet]

I believe that......

And I don't. Seeing as we're both not Google, our opinions on this topic are pretty much moot.

Yeah, I guess we should just take this off to some blog/discussion site then ...

Google want to sell more reCaptcha

[Anonymous Coward]

Google has seen an opportunity to extend reCaptcha to phone services. They create robot calls and therefore create demand for a system to verify that callers are not robots. So ironic that a company dependent on computing indulges so many anti-computer technologies.

Hangups

[sgunhouse]

I don't know about you, but if I get a call from an automated system I just hang up. If the call starts off immediately by saying it is automated I'm sure that is what will happen. The first thing that has to happen is to indicate to the recipient what the call is about; after that they can say (especially if there is a response the system doesn't understand) that they are a machine.

Re:Hangups

[93 Escort Wagon]

I don't know about you, but if I get a call from an automated system I just hang up.

As do I. But I still feel as if there should be an easy and obvious way for me to make that determination.

At first thought, anyway, this "Duplex" thing rather annoys me. If the "person" on whose behalf the bot is calling doesn't feel it is worth their time to speak to me directly, why should I have to waste my time talking to their bot?

Re:Hangups

[LetterRip]

At first thought, anyway, this "Duplex" thing rather annoys me. If the "person" on whose behalf the bot is calling doesn't feel it is worth their time to speak to me directly, why should I have to waste my time talking to their bot?

This bot makes reservations. As an employee of the company that the reservation is being made at, you wouldn't be "wasting your time" - you would be doing your job.

Re:Hangups

[Guillermito]

if I get a call from an automated system I just hang up

This is because the vast majority of calls we get from automated systems now are unsolicited and aimed at parting us with our money. On the other hand, businesses are eager to get the kind of calls that Google demoed. Plus the robot seemed very polite and down to business. I'm pretty sure client facing employees would love to talk to the Google AI bot instead of dealing with rude, incoherent, indecisive human customers.

Re:Hangups

[stephanruby]

I completely agree. But there are times where Duplex could be useful.

"Ok Google, call the radio show. I need to be the 99th caller to win.", "Ok Google, buy me this ticket for this show before it gets sold out.", "Ok Google, call Xfinity, pretend that I want to move to AT&T unless they cut me a new discounted rate", "Ok Google, pretend you're an elderly woman and waste as much of this scammer's time as possible.", "Ok Google, please pick up whenever my mother-in-law calls. Tell her I am busy. Ask her what she wants and send me a summary of her complaints by SMS."

Re:Hangups

[POPE Mad Mitch]

If i get a call from what is obviously a call center i just want to hang up, they always start with phony platitudes, asking how am I today etc. I don't know you, anything I could possibly say is meaningless to you, why are you wasting my time even bothering, just tell me what the f. you want and go away. I would so much prefer an automated system if that didnt try to be my friend and just got the hell on with it. Also wouldnt have the background roar of hundreds of other conversations happening in the same room as the caller.

Re:Hangups

[bazorg]

But there are times where Duplex could be useful.

Perhaps that guy from Pulp Fiction who had a speech impediment and still wants to order pizza from a place that does not do ecommerce.

Plenty of other scenarios can be imagined, involving people who work 9-5 and cannot be in the phone queue for some service that is also only open 9-5.

This can also be seen as a good way to avoid dodgy upselling. My robot is calling for the offer X that was advertised and your selling bot can argue for hours that there's a better deal, but I won't take it.

Re:Hangups

[bazorg]

OH! and I nearly forgot... Joe just had a car crash and needs to ring the emergency services and give them precise location... much better to have a bot figure those things out while the owner is busy trying not to to bleed to death.

Re:Hangups

[EvilSS]

I don't know about you, but if I get a call from an automated system I just hang up.

As do I. But I still feel as if there should be an easy and obvious way for me to make that determination.

At first thought, anyway, this "Duplex" thing rather annoys me. If the "person" on whose behalf the bot is calling doesn't feel it is worth their time to speak to me directly, why should I have to waste my time talking to their bot?

Because you work at a business that wants their money? If you own the business and don't want their money, fine. If you are an employee, suck it up and do as you are told.

Re:Hangups

[squiggleslash]

OK but are you a restaurant that accepts reservations? And if so, if the vast majority of automated calls were from people making reservations, wouldn't that be fucking stupid of you?

Re:Hangups

[david_thornley]

If my car detects crashing, and doesn't get instructions from me in a certain period of time, it will attempt to call emergency services and give them my location.

cut out the people

[petes_PoV]

With luck, soon both sides: the appointment-making and the reservation-taking will be given over to the machines. So this will simply be my Duplex calling your Duplex. I can see some benefits to each of them knowing they are talking to (essentially) themself, that way they could both hang up and negotiate whatever the call was about far more efficiently in a few milliseconds.

It is only while there is the possibility that one system is so archaic that it still has an actual person taking the call that there is a difficulty. But even then, it's not much of an issue, what with the Duplex system being backwards compatible with meat.

Re:cut out the people

[Dutch Gun]

If a Duplex realizes it's talking to another Duplex, will it go into 56K modem mode and just talk electronically?

Re:cut out the people

[guruevi]

Why not, its more efficient. This could be as easy as a nearly inaudible click added when the line is opened.

The alternate option, when this becomes wider deployed is that the thing develops its own dialect through machine learning.

Re:cut out the people

[locofungus]

With luck, soon both sides: the appointment-making and the reservation-taking will be given over to the machines.

That would be awesome but I can't see it happening for a long time.

Hell, *I* cannot navigate the menus to get to talk to a person. What hope has the google AI.

Most recently I needed to change the *mailing* address for my council tax bill (think property taxes)

Now it's slightly unusual to have a mailing address different to the property address but sufficiently common that all council telephone line staff will have dealt with it dozens of times before.

So I ring up. 'this is the 24 hour... During the hours 9am to 4pm you will also be able to talk to a person'

So far, so good.

Now none of the options quite fit - am I moving home or not? So I guess, next option 'hmmm, I think I guessed wrong on the previous option but I'll get to a person eventually'. Anyway, eventually I get to a 'you can find the forms at www.blurb. Goodbye'

Try again. Different route, same outcome. Try again - Do nothing 'we did not detect a key press. Goodbye'

Eventually, after 5 or six attempts, I managed to get through to a person and 30 seconds later was done with the call.

And it will get worse. As AI improves, the number of calls that it will refuse to route to a person will increase making it harder and harder to actually deal with these systems. I never call them unless I cannot work out how to deal with the question without a call.

Re:cut out the people

[POPE Mad Mitch]

They could cut out the people by just having an online booking service in the first place, either their own, or joining one of the many 3rd party services. It is purely because they haven't done this that duplex needs to exist in the first place. So if the receiving side ever got automated enough so as to have their own duplex call answering then google wouldn't need to call them in the first place.

Re:cut out the people

[Ksevio]

I imagine it would go like the pods in Nier Automata

Warning? This almost passes the Turing test

[Anonymous Coward]

Why would you spoil a perfectly good turing-test with a warning?

Re:The real question is...

[guruevi]

People get what they want, easier and faster.

Getting this ready for telemarketers

[Anonymous Coward]

I want this ready to keep up a conversation with a telemarketer.

Demonstrates misunderstanding of AI

[aberglas]

It is very difficult to know whether these machines are intelligent or not. If they are just playing to some fairly fixed scripts then as long as the person at the other end stays within the script then no real intelligence is needed. Eliza/Doctor did this sort of thing 50 years ago by simple pattern recognition on sentences.

Sure, this system is smarter than Eliza (hopefully), but I suspect that the moment you go off script it fails catastrophically. The human would soon tweak that they were talking to an automated bot (even if they were actually talking to another human that was not too smart!).

These things have the potential to be really annoying.

Eventually, they may know what a restaurant booking really is beyond the superficial words and phrases. At that point people will be redundant. But that is still decades away.

Re:Meh

[Anonymous Coward]

According to the latest LLVM debacle, yes. White males like me are indeed being specifically singled out, targeted, and attacked. It is hateful, racist and has as much place in our society as any other gender or racial bias.

Or are you saying it is okay to discriminate against me? Because that is the vibe I am getting from google and many others. I have never discriminated against anyone before and I'll be damned if I will let it happen to me.

Your not the first one to say that I am 'crying' because I am white, *hint* your still a racist asshole no matter who your being racist against.

Figuring it is A-OKAY because I am white just proves my point.

google's profit model with this

[SethJohnson]

I strongly suspect that Google is planning to lease services to businesses that will field these duplex-sourced calls without routing them to their human staff. Google has an opportunity to generate a bunch of annoying bot traffic to human call-takers and then sell the solution to the businesses to automate the transactions. Instead of one bot talking to another bot, they'll probably pass JSONs back and forth to negotiate whatever it is the Google Assistant is attempting to arrange.

Observer Effect

[mykro76]

That would defeat the purpose of training the AI on genuine human conversations. If the recipient knew it was an AI calling they would be likely to change their behaviour such as talking in shorter, simpler sentences with overly exaggerated pronunciation.

Re:Observer Effect

[Actually, I do RTFA]

And maybe Google shouldn't be allowed to do something just because it helps their AI development? Certainly, if I'm putting forth effort in generating their corpus, I want compensation.

Ya, no.

[fahrbot-bot]

... people are just fine with talking to automated systems so long as they are aware of the fact that they are not talking to another person ...

Perhaps this isn't the same thing, but ... I much prefer pressing buttons than "talking" to an automated system. It's simpler and more private. (Why would I want to say things that could be overheard -- like in that TV commercial where a guy is saying his credit card number aloud for the automated system... and I'd much rather "Press 1 to speak to the Proctologist" than say "Proctologist" ...) Sure voice systems may allow more varied options and interactions than what can be easily be supported by pressing buttons, but it's super annoying to talk at the phone. Also, I like to reserve my irate vocalizations for my PCs running Windows.

Lastly, I hate systems that combine pressing numbers *and* requiring voice inputs at different, seemingly random, times during the call -- grrr...

Re:Ya, no.

[omnichad]

Inevitably, these systems ask for the account number verbally, but never hint at the fact that you can key in the number instead - even though it does work.

0 != O

[WaxParadigm]

Don't know about the announcement, but after listening to the calls I think someone should teach Duplex that round digits in phone numbers are zeros (not the letter o).

Re:0 != O

[RyoShin]

Except that "Oh" is a common (American) way to say zero when saying a number by individual digit. Think 404 (Four-oh-four); at least, I've never heard anyone say "four-zero-four". So if Duplex is trying to imitate speech patterns, it's proper for it to say the number that way in at least some circumstances.

(When I was in BCT there were more than a few moments of drill sergeants yelling at a recruit "What fucking number does 'Oh' come after?")

abuse

[Tom]

I'm much less concerned about the "omg the person I was talking to wasn't actually a person" and much more about the abuse potential of the whole thing.

A lot of the real world works because as human beings we can generally trust each other, exceptions are rare enough to not break the system and a personal interaction establishes a slim line of bidirectional trust.

If you have access to such a system, and I'm certain they will make it generally available, there's a business there, you can now flood the restaurant or hairdresser etc you didn't like with fake reservations, denying them actual business.

They will have to answer with verification systems, which a) makes everything more complicated for us actual humans and b) adds a small overhead and c) just starts the arms race we already know from IT security.

And that is just the very first thing that comes to mind. Criminals are sure to be more creative than that. These systems are disruptive, and I haven't seen anyone thinking about solutions to that so far. Maybe the world after we solve this will be better than the world now, but it will be a major change. I'm reasonably sure that reservations of all kinds via telephone will go away. When the dust settles, you will no longer call a restaurant to ask for a table. You will tell your smartphone to reserve one, which will then call the restaurants computer, they will manage the verification and validation details in the background and generate a token that you can show at the entrance to get your table.

I'm a tech person, I feel comfortable with that. I would probably prefer it over calling the restaurant and speaking to a real person and we barely understand each other because of the noise in the background, etc. - but many people prefer to actually interact with an actual human being and that will be lost to them.

Re:abuse

[swillden]

If you have access to such a system, and I'm certain they will make it generally available... These systems are disruptive, and I haven't seen anyone thinking about solutions to that so far.

You are assuming that no one has thought about it, and that assumption makes you assume that Google will make this generally available. It's far more likely that Google will not make it generally available because they have thought about it. It's just as saleable as a service (perhaps more profitably, actually), and as a service Google can take steps to minimize abuse.

They will have to answer with verification systems, which a) makes everything more complicated for us actual humans and b) adds a small overhead and c) just starts the arms race we already know from IT security.

Even if this did happen, it wouldn't be an "arms race". The verification would be a credit card number, and the restaurant would hold a reservation fee which wouldn't be completed unless you didn't show (or could be fully charged at point of reservation and deducted from your bill when you eat). This would indeed be more complicated and add a small overhead, but no arms race. Restaurants would set the fee high enough to deter fake reservations, and that would be the end of it. It wouldn't be unprecedented, either, because some high-end restaurants already do it.

Re:abuse

[david_thornley]

Prank and malicious phone calls have existed for a long time.

Re:abuse

[Tom]

and as a service Google can take steps to minimize abuse.

Which steps would that be?

We haven't been able to solve the issue of spam, despite trying for 25 years. What makes you think those mysterious "steps" would be any more successful?

The verification would be a credit card number

a) check
b) check
c) well, since you absolutely can't get such a number anywhere unless you are the rightful owner of said card...

Re:abuse

[Tom]

Yes, but you could not automated them and spam them at the rate of your outgoing bandwidth. That is the point.

Re:abuse

[swillden]

and as a service Google can take steps to minimize abuse.

Which steps would that be?

Oh, maybe noticing when a business is being flooded with calls and refusing to make them?

We haven't been able to solve the issue of spam, despite trying for 25 years. What makes you think those mysterious "steps" would be any more successful?

The thing that makes spam hard to address is the decentralized nature of SMTP. If all email originated from a single system it would be easy.

The verification would be a credit card number

a) check b) check c) well, since you absolutely can't get such a number anywhere unless you are the rightful owner of said card...

And yet this isn't a significant problem for hotels and restaurants that already do this. The exact reasons it's not a significant problem are complex. If you really want a lecture on how the credit card fraud mitigation systems work, I could provide it, but I think it's adequate to point out that if the issue you're positing were real, it would already be a bigger problem than it is. (Which isn't to say that it's not a problem, but it's a manageable one.)

Re:abuse

[Tom]

Oh, maybe noticing when a business is being flooded with calls and refusing to make them?

Congratulations, you've just replaced one DoS attack with a different DoS attack.

The thing that makes spam hard to address is the decentralized nature of SMTP. If all email originated from a single system it would be easy.

What makes you think these systems would be centralized?

And yet this isn't a significant problem for hotels and restaurants that already do this.

Some hotels do, that is true. A small minority of restaurants do. I'm not aware of any hairdressers that ask for your credit card on the phone.

DOA

[duke_cheetah2003]

I believe that all production environment calls (essentially, calls not being made for internal test purposes) from Google's Duplex system should be required by Google to include an initial verbal warning to the called party that they have been called by an automated system, not by a human being -- the exact wording of that announcement to be determined.

OK First this right here is a major hurdle. If you make this thing warn people it's a robocall, most folks will just hang up immediately, thinking it's yet another sales pitch, or free cruise, or health insurance, gawd there's so many now. So this is DOA if it's gotta announce it's automated.

But as a side note, it's going to be amusing when our 'AI's start calling each other to whatever, communicating in simulated english or whatever spoken language the systems in question are trained on.

Obviously the real solution is for your hair stylist and favorite restaurant have some non-verbal mechanism for arranging appointments or reservations or whatever. Obviously every one and every company and every little this and that can't have their own App for achieving this, which is what some larger companies are doing to migrate 'ordering stuff' from humans talking to humans to humans just fondling their portable personal computer.

I'm imagine whomever cobbles such a system together and convinces a large segment of the population to use it is going to be rich. But this Duplex thing? DOA. The stigma surrounding robocalls is all bad, and all deserved I'm afraid.

Re:DOA

[misexistentialist]

Large companies have been using computers for basic phone tasks for a while, if there is a stigma isn't doesn't matter because you have no choice

Re:DOA

[duke_cheetah2003]

Large companies have been using computers for basic phone tasks for a while, if there is a stigma isn't doesn't matter because you have no choice

Completely different. I'm assuming you're speaking of voicemail in all it's incarnations. Totally different. You are calling a robot. And you know you are, and who you're calling. When a robot calls you, at the moment, it's almost always telemarketers, scammers, or surveyors. And most people don't want to talk to those kinds of robots. To the point, as soon as they conclude it's a robot, they hang-up.

In my case, I even hang-up on robots I solicit calls from. Like my pharmacy, they have an automated system that calls when scripts are ready, and as soon as I realize it's them, I just hang up and proceed to the pharmacy. Nothing further I need to know beyond it's ready.

Leaching

[Anonymous Coward]

My feeling (this is my feeling and I lay claim to it) is that as technology progresses, the humanity gets leached out of everyday life. People hide from having to talk to a human being by getting Duplex to talk to a human being. Ultimately that other human being feels dehumanised as they take requests from a Duplex. And further, that human being ultimately gets replaced by another Duplex.

Do we have to automate everything? Social interaction is fundamental to our wellbeing as a society.

Perhaps I'm just getting old - but I wonder if the tech evangelists even think about societal impact.

robots

[ico2]

I imagine it'll go much like when I call the bank and get a robot: "Please tell me what you are calling about today" "speak to operator" "sorry, I didn't catch that, you can use phrases such as 'check my balance' or 'order a replacement card'" "speak to operator" "sorry, I..." "SPEAK TO OPERATOR" "sorr..." "FUCK OFF" "sor..." "FUCK OFF FUCK OFF FUCK OFF" "I'll connect you to someone who can help"

Opinion from a Googler

[Anonymous Coward]

I asked a Google presenter at an IO Extended event a similar question - namely, that as a user of Google products, I have consented to interact with Google's software, but someone receiving a call from Google's AI has not consented to being called by a bot.

Their (personal) opinion was that this isn't an issue. A business with a listed phone number that invites calls for the specific purpose of booking appointments/reservations WANTS to receive calls for and on behalf of customers. There's no difference between Google doing a call and a concierge doing a the same call.

I mentioned that I felt there should be some kind of warning, or that the business owner might give a one-time consent even if the front-line employees were unaware, but they stuck to their opinion.

As an aside, it was refreshing that the Google employee was free to share his personal opinion, as he did for several other questions, even when he disagreed with Google's direction.

The next step

[garryknight]

"Siri, answer the phone."

Suggested "warning" message

[shayd2]

Hello, this is one of your masters.

Everyone is worried about the service staff

[SmSlDoo]

Just wait until Comcast or some other major sp(c)ammer gets their grubby mitts on this.
Expect to get phone calls every day as they want to sell you the latest and greatest crap you don't want.

Need to get your internet fixed? Why pay a worker in a call center $0.03/hour to take your call when you can just have duplex do it?

Re:Everyone is worried about the service staff

[EvilSS]

Need to get your internet fixed? Why pay a worker in a call center $0.03/hour to take your call when you can just have duplex do it?

That would be a massive improvement in customer service if they did.

No

[nospam007]

A Robocalls is a Robocall is a Robocall.

Hand picked examples

[sjbe]

Why should they? There is no logical reason for them to do so. If the bot works as well in reality as it did in the three demos, thern there is no reason to 'warn' the person on the other end that it is a bot.

Are you aware of the concepts of selection bias and survivorship bias? These are hand picked demos intended to make the technology look as good as possible. I'm deeply dubious it would perform as well under real world conditions.

While I actually prefer not having to talk to a person in a lot of cases, I've never seen a machine or program that could even come close to properly interpreting my requests using voice commands on a reliable basis outside of a few narrow use cases. Presumably this system is better than Siri and equivalents but that's a pretty low bar to set. And if I'm not talking to a human I want to know that fact up front so I can adjust my actions accordingly. Odds are I would figure it out quickly enough anyway but it should be disclosed.

No show ?

[DrYak]

Example: let's say you're going to one of those restaurants where you pay a 'penalty' for no show (but you, the person instructing the Duplex don't know this); the human mentions this to the Duplex. Is this a valid contract? Who knows?

How does a business that makes you pay a penalty for no show would *actually make you pay* the penalty *when you don't show* ?!
I'm genuinely interested ?

Do they request your credit card information at the moment of the table reservation, the same as hotels with similar penalty do ?
Do they ask for these it *over the phone* ? (remember, the whole purpose of Google Duplex is to give a computer-to-human interface for booking business that still lack any modern reservations system and still rely solely on a human answering the phone)
That would be a gigantic security failure.
(And even technical impossibility: lots of modern European Chip-and-PIN credit card cannot be billed without a second form of confirmation. If it's not a terminal asking for a pin code, nor a website implementing 3DSecure - then either the credit card company calls you directly to confirm or you need to use the card's smartphone app to confirm the transaction out-of-band)

(Which, by the way gives a clear path google :
- simply refuse to book restaurants that ask credit cards information on the phone. Which again, makes sens from a security and technical point of view.
or alternatively :
- give a special credit card number which is operated by google, and insured against fraud and even insured against "no show" as part as a paid-for service to monetize Duplex)

Re:Why a warning? Robocalls are illegal.

[kent.dickey]

I agree, this would appear to be an illegal robocall. I don't understand how Google doesn't realize this is a problem.

For those in favor of this, what would you think of v2.0 of the appointment robot automatically calling ALL restaurants within a 5 mile radius to book a table? And there's no reason all the calls cannot be simultaneous. And then calling back ALL-1 to cancel. Restaurants will need to hire a team to staff their phones (or, just automate it as well...this may be what Google has in mind, force businesses to automate their phones as well). And then v3.0 is competitors abuse the system: do something to get users of this product to Denial-of-Service their competitors. It's obvious that any automated call system is ripe for abuse by unsavory actors by tricking normal people to do something abusive.

And, lowly workers now have to grovel at the sounds of the automated butlers of the rich. I would expect a backlash.

Obvious Google baloney

[JustNiz]

> With no exceptions so far, the sense of these reactions has confirmed what I suspected -- that people are just fine with talking to automated systems so long as they are aware of the fact that they are not talking to another person.

Let me be the first then. I hate the damn things, and I REALLY can't believe that I am exceptional in that respect.

Re:Obvious Google baloney

[brewthatistrue]

Agreed. People hate phone trees as well.

https://gethuman.com/

Robocallers?

[WillyWanker]

And when will this apply to robocallers, spammers, and debt collectors who use pre-recorded messages that try to trick people into thinking a human is calling them? Y'know, shit that they've been doing for over a decade and without anyone even raising an eyebrow.

If Google is going to be required to disclose it's an AI call, then all robocallers should have to as well.

Polite to say who you are.

[gurps_npc]

They should introduce themselves like this:

"Hello, this is Google's Duplex, calling for John Doe. John would like a table for two at 6 PM this Saturday....."

Meanwhile, in the real world

[lamer01]

My city still expects me to show up in town hall with my car registration document in order to get a beach/park permit.

Re:Meanwhile, in the real world

[omnichad]

You're required to own a car for what now?

No announcement necessary

[Tactical Bacon]

There's no need to announce it's an AI. Anyone who's worked phones for customer service before can tell you that when a call is handled quickly and efficiently odds are pretty good it wasn't an actual human. The lack of whiny insistence that they get an appointment during a time that's not available, asking the same question in 3 different ways, and otherwise ignoring what the person helping you is saying will be immediate clues that it's an AI. I would *love* if all the people that call me for tech support would let an AI handle the call instead; it would eliminate a whole lot of frustration and wasted time.

"Hello! I would like to make a reservation!"

[tgibson]

"Press 1 if you would like me to provide the date first, press 2 if you would like me to provide the time of day first. Presiona 3 para escuchar esto en español"

(beep)

"Thank you. You pressed 2 for the time of day first. If this is correct, please press 1."

(beep)
"Thank you. Please press 1 if you would like the time in 24 hour clock notation. Please press 2 if you would like the time in AM/PM notation."

(boop)

"I'm sorry. That is not a valid option. Please stay on the line and a reservation-maker will be with you shortly."

good robocalls

[trb]

Opinion 1: Google Duplex is a robocall.
Opinion 2: But it's a good robocall!
Duplex: Hello, this is a good robocall.
Every other losing robocall: Hello, this is a good robocall.

This could result in less work for them

[grahamsz]

Ultimately if google calls up every store in the country and asks "Are you open on your normal hours on memorial day?" and can then put that data into their search results and confirm that they've checked the opening hours then that'll mean less random people calling up to ask the same question.

Same thing when a storm hits, google could call and get updates about whether a business plans to close early and relay that information to the rest of us.

I suppose the problem will be if twenty different companies call to get the exact same information.

Re:The real question is...

[Sloppy]

If you want it easier and faster, then just get rid of the phones and have the processes connect to sockets and perform RFCwhatever appointment-making protocol.

I would refuse to deal with it

[Rick Schumann]

If I were running a restaurant or business and I got an 'automated' call from something like this, I'd hang up on it. If it called back I might try demanding to speak to an actual human being, but if that didn't work then I'd just keep hanging up on it. Why, you ask? Because I'd have no way of knowing for sure that the call was legitimate, not some sort of prank, and not the result of some malfunction, unintended activation, or someone hacking someone else's hardware. I'd insist on verification by an actual human being, and I don't think I'm alone on that, regardless of anything Google might have to say about their 'test calls'.

Re:Meh

[squiggleslash]

Fun fact: when a survey was done about people's attitudes towards diversity and their skills, the graph was almost a straight line between "Skilled, in favor", and "Mediocre, scared of diversity".

People who have valuable skills aren't scared that they're going to be replaced by a one legged black lesbian transwoman.

Illegal Recording?

[TexasTroy]

As soon as someone responds, the bot is writing the conversation to memory. How is this not making an illegal recording, regardless of length of storage, in two party consent locations?

Re:Meh

[david_thornley]

Nope. What LLVM was doing was working with a place that recommends candidates who are neither white men nor Asian men. LLVM didn't actually assign an intern slot to that place. In the meantime, you enjoy privilege as a white male. You're more likely to be taken seriously. You're more likely to have your resume picked out for an interview. There's plenty of other ways you benefit. I don't see you complaining about those.

However, let there be some extra consideration for people that don't look like you, even when you have an advantage in less formal ways, and you whine.

I'm not claiming that you're crying because you are white. I'm saying you don't know what you're talking about, and you're crying because you don't want to give up any unfair advantages you've got.

Re:Why a warning? Robocalls are illegal.

[david_thornley]

I've got one phone line coming into my house. Either it is busy, or it isn't. If it's busy, it can't be used for an outgoing call. Therefore, the computer can't make such calls significantly faster than I can (it will presumably dial faster, so it may be slightly faster). It does take much of the work out of prank phone calls, but I don't remember the people I knew who were into prank phone calls trying to avoid the hypothetical drudgery.

Re:The real question is...

[guruevi]

Obviously that would be ideal but while everyone is switching over from "having to man the phones" to "let the robot figure it out" there is going to be some transition.