Aardvark DailyNew Zealand's longest-running online daily news and commentary publication, now in its 24th year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.
Content copyright © 1995 - 2018 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk
Please visit the sponsor!
I'm sure all Aardvark readers have already heard the recorded conversation between a hair-dressing salon receptionist and Google's AI "bot". If not, go search for it and listen.
People have been in awe of this demonstration which has the most incredibly life-like synthesized voice and algorithms that interpret what's being said by the human end of the conversation with 100% accuracy.
My own experiences with voice recognition, even when a computer is trying to cope with structured prose or single-word responses, is nowhere near as good as that supposedly demonstrated by Google in this demo.
So should we be in awe -- or should we be suspicious that this was a "smoke and mirrors" demo, somewhat akin to Kim Jong Un's 3D-rendered military might?
Well a number of people who are in a position to make critical comment have come out claiming that they're not convinced that this demo was real.
As *we* all know, marketing is about creating a perception, an illusion, a belief -- not necessarily about reality.
Some of the key flaws pointed out by people smarter than myself include the fact that the salon receptionist does not identify the business when she takes the call. Almost without exception, business phones are answered by stating the name of the business. Indeed, in an ad-hoc survey, 100% of all businesses rung by one skeptic included the name of the business in their phone greeting when answering such calls.
Then there's the small, but *very* important fact that it is illegal to record phone conversations in the state of California without the permission of all the parties involved.
Surely Google didn't break the law to make this demo?
Even if they obtained the permission of the salon receptionist after the call was made, it was still an illegal act to record it without prior permission. I wonder if they'll be prosecuted for this -- or have they already admitted to law-enforcement that this was just a mock-up?
But what if it was all kosher and above-board?
Should we be worried that AI and voice synth/recognition systems are now seemingly so good that they are indistinguishable from a real human being?
It's clear that the "bot" used by Google in its demo would pass the Touring Test with flying colours and would fool virtually anyone who was the other party in such a conversation. What does that mean for the future?
Will those rooms filled with young Indian people scamming us out of our hard earned cash by claiming our computers are infected with "worms and wiruses" soon be replaced with just a room full of computers -- that need no pay, no food, no water?
Will the mundane human-interface roles (such as receptionists, phone operators, etc) be replaced by "bots"?
Remember the recent claims that AI will take-over a huge chunk of jobs currently performed by humans? This seems to validate that prediction somewhat.
Perhaps, in just a few short years, the combination of self-driving cars and human-like AI will mean that Johnny-Cabs will move from the realm of scifi to reality.
Only time will tell -- but place your bets now!
Please visit the sponsor!
Have your say in the Aardvark Forums.