Google wanted to make its voice
assistance sound more human but the
company may have taken things a step too
far
Google revealed that it is able to make
a computer speak in a way that can fool
others into thinking it is human just
listen in to this example of the Google
assistant robot calling a hair salon
receptionist to book an appointment on
behalf of its human owner now happening
healthier hi I'm calling to book a
woman's haircut for our clients um I'm
looking for something on May 3rd
sure give me one second
mm-hmm so our what time are you looking
for a loan at 12:00 p.m. we do not have
a 12 p.m. available the closest we have
to that is a 1:15 do you have anything
between 10:00 a.m. and 12:00 p.m.
depending on what service she would like
what service is she looking for just a
woman's haircut for now ok we have a 10
o'clock 10:00 a.m. is fine
not only did Google eliminate the
typical robot sounding voice but the
computer is using language fillers and
quirks to make it sound more real there
were owns and moons and the computer
even used upspeak which is that trend
where people speak phrases with an
upward Bend so it sounds like a question
Google calls this new assistant
technology duplex and the plan is to
launch this advanced artificial
intelligence in the summer with limited
release now here's how it would work you
pull out your phone ask the Google Voice
assistant to book you a reservation
somewhere then the robot on its own goes
to work if it cannot make the
reservation electronically say using
something like open table it will jump
on the phone and call the business to
handle the back-and-forth convo of
making your appointment you will get an
alert when it's done you won't even hear
the conversation that took place here's
another example Google provided this
time the robot assistant is a male voice
looking a restaurant reservation over
the phone
hi I'd like to reserve a table for
Wednesday the set
oh they win people it's for four people
four people win next Wednesday at 6 p.m.
Oh actually the leaves are for light
opera like apply people for beautiful
before you can come how long is the way
usually to be seated
when tomorrow all we care for next
Wednesday the 7th oh no it's not too
easy you can come from Home Depot
okay oh I got you
banging once again Google can keep up
with complex questions it sounds super
smooth with realistic mannerisms like oh
I gotcha and there's nothing to indicate
you're talking to something that is not
alive and that's a problem if technology
is at the point where we cannot tell
what is a real person and what is a
robot then we need to start talking
about the ethics here should you have
the right to know when you're talking to
a robot does Google have a
responsibility to disclose when its
artificial intelligence is posing as a
human we are living in an age of
disinformation right now people can't
tell the difference between made up
garbage on the Internet and real news
photos are manipulated to play off of
our emotions and fool us on social media
and now you can't even trust your own
ears Google is showing us how easy it is
for dialogue to be spoofed what can we
believe anymore
I should note Google has said that when
the duplex bot launches it will likely
tell the person on the other end that it
is a digital assistant but that's a
disclosure not present in any of their
example phone calls and it's not clear
how Google would try to weave that into
the conversation making hair
appointments and booking tables at
restaurants may well be harmless and
even helpful but without disclosure the
core feature of duplex is in my opinion
a deception in this case vocal deception
why not just make it a smooth sounding
voice why add the imperfections like
rooms or us google says these phrases
can help
the pauses the computer would need while
it's processing the information so that
means Google took extra steps to
disguise the robot to help hide the
machine running in the background why
even go that far if you also want to be
transparent as Google says it will be
why are we creating things to trick
humans I'd hate to think of this
technology getting in the hands of
scammers making robo calls preying on
the vulnerable yes those problems
already exists before duplex but imagine
if Kahneman can now operate at scale
with friendly back-and-forth banter and
what about those political phone calls
trying to sway your vote
it'll be increasingly harder to tell
what is real Google may be proud of how
humans can't tell when they are talking
to a machine but while Google was
celebrating did anyone stop to think
what it means for society before making
technology is anyone asking questions of
if we should make it because if Google
or another company uses this without
transparency then we'll be the ones left
asking a question every time we pick up
the phone hi are you alive and if the
voice on the other end stumbles before
they answer you still may not know
thanks for watching let me know in the
comments if you agree disagree maybe you
see this as the first step in a cool
future when we could just talk to our
gadgets and these are issues that are
gonna get worked out or maybe you're
also feeling the creeps with this audio
uncanny valley zone that Google just
entered be sure to subscribe and click
on that Bell icon to keep up with all
the biggest stories coming out of Google
i/o this week
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.