Through the looking glass

July 2018

I can’t help feeling that if we are not actually through the looking glass, we are well on the way.  Google’s latest invention is at the same time impressive and chilling.  They have found a way for a computer to make telephone calls on your behalf and for the person receiving the call to be completely unaware that they are talking to a computer.

Google already sells a gadget called Google Assistant (Amazon have the similar thing called Alexa), which answers your verbal questions and performs basic tasks if instructed.  It can tell you the time, the weather, what’s in your diary, even switch the lights on if you’ve set it up to do so.  So far so good.

However, Google Assistant, like Alexa, has hitherto been a meek servant, only speaking when spoken to, obeying orders without complaint.

No more; at a recent conference Google demonstrated a new capability that will be placed in Google Assistant shortly.  It is called Duplex and turns Google Assistant proactive, allowing it to make phone calls that sound so convincing that you would swear it’s a real person.

Google showed this off at a recent conference; the computer called a hair dresser to make an appointment for a haircut.  The most unsettling part was the human intonation that they had built into it; ums and ahs, pauses, use of idiom, saying ‘mm-hmm’ rather than ‘yes’.  It had to make some decisions based on what the person being called said, and it managed that well.  As technology, it’s astonishing; to that extent I take my hat off to them.  You can see it yourself below this article.

So we now have a machine that can impersonate a human being convincingly; this is not a ‘press one for sales’ sort of thing, this actually seems to be a person, which reacts to what it hears.  Humans stumble over words, pause to consider a phrase, correct themselves – computers don’t, or at least used not to.

Technological brilliance aside, it’s the ethical issues that threaten to lead us through the looking glass.  The prospect for abuse is considerable; scam telephone calls are bad enough when there is a person making them; just imagine how much more frequent, and probably more convincing, they will be if made by this technology. 

To be fair to the Google people, they have said that they will develop a mechanism to allow us to know that we are talking to a robot.  Perhaps they will, but let’s not forget that Google is certainly not the only company working on this; it’s just the only one being open about it.  Goodness knows how many shadowy but brilliant laboratories all over the world are heading down the same road.

I have no doubt that soon we will have many providers of this form of artificial intelligence that can sell to you, cajole you, threaten you, scare you, reassure you, inform you, warn you and more. Governments will love it, as will crooks.

So, who should be making sure that we are protected from abuse of this technology?  I’ll tell you who; it’s us.  We can’t trust the likes of Google to act in our best interests, whatever they say, and it is for us to impose some disciple and boundaries on them to make sure that this impressive development is used wisely and for our benefit.  The trouble is, I don’t know how we are going to do it.

We can’t un-invent the machine, nor should we want to; I have no doubt that many admirable uses will be found for it.  However, whilst we may not be completely through the looking glass just yet, as Dorothy said to Toto: I've a feeling we're not in Kansas anymore.

____________________________________

Videos

Above is Google’s eerie demonstration of the phone call. 

___________________________________________________________________

If you really get the bit between your teeth, you can watch the whole three day conference by clicking here.

____________________________________________________________________________

On the other hand, here is one robot dog opening the door for another – such good manners:

Meanwhile, here is comedian Michael McIntyre making much more sense: