let me take you back to 1960 Mattel has
just launched Barbie the madmen are
revolutionising car advertising and you
can still smoke in your office if you
work at a brand in 1960 and you’re
trying to create the next greatest
product and you’ve got to be able to
predict where your customers want so you
go out and talk to them you stand in a
mall and hand out questionnaires you run
a focus group you do market research
fast-forward 60 years later to today and
companies still need to talk to their
customers and they still do it in
exactly the same way so think of the
things that you bought this week I got
some new clothes pair of headphones a
can of coke almost all the things that
you bought would have been created out
of our discussion with small groups of
people using pretty much the same
techniques as 60 years ago and that
isn’t because they work they frequently
don’t the reality is the vast majority
of new products fail the wrong things
just get built and so we live in a world
that is biased away from innovation
towards promotion last year we spent
five hundred billion dollars on
advertising and that’s money that is
being used to convince us what to buy
but that does next to nothing to
actually improve the product and
companies are constantly spamming us
telling us what we want and it works
you’re actually looking at a man who now
buys more expensive shampoo than he did
when he had a full head of hair
that is a true story so what if the
world was less wasteful what if we could
create a future where all the energy
that we currently spend on advertising
was instead spent on making a better
product a future where companies were so
good at listening to us that their
innovation replace the need for
advertising altogether in that future
marketing isn’t about pushing and
promotion it’s about managing the flow
of information between those that make
products and those that consume products
now I believe that this feature is
possible and the key is our relationship
with artificial intelligence at the very
core of AI we’re trying to create
machines that learn machines that learn
in many ways like you and I do and in
fact one of the main things that these
machines that need to learn about is us
but whilst you and I are pretty limited
on the number of people that we can
learn about at the same time the
machines that we’re designing can
observe and interact with millions of
people simultaneously they can combine
that knowledge and form very accurate
models about how each of us behaves and
there are plenty of these machines
already on the internet constantly
learning about us watching what we click
on what we buy who we connect with and
the longer they watch the better they
get at predicting where we might click
next what we might like to buy who our
friends might be we’re used to this
whole idea of what we’ve now become has
become known as personalization highly
tuned adverse chases around the internet
however the products they are selling
don’t get better and so we’re back to
where we were in 1960 sure we’re far
more connected now but our personal data
is being used in a way that we can’t
readily track to form incredibly
sophisticated models of our purchasing
behavior and those models are then used
to convince us to buy more things and
they still do not improve the products
that which one are creates creates a lot
of value for businesses but nowhere near
enough value for us consumers
so last year we said about trying to
rethink this we wanted to see if in the
future the most effective brands are the
ones that talk to their customers not to
make better advertising but to share
ideas we thought that they’d use
artificial intelligence to help real
people and brands talk to each other and
we wanted to prove this in an industry
which people have very strong views on
and that which we had a pretty
significant interest in as well
beer so we created intelligent X the
world’s first beer brewed by AI by
incorporating AI into the brewing
process we have created a product that
automatically changes its recipe based
on the needs of the consumers a beer
that continually adapts to the changing
tastes of the people drinking it now
that is what I call personalization and
it works very simply on one side
algorithms are deciding which
ingredients to use how long’s to cook
things for and on the other they’re
asking questions for the people actually
drinking the beer questions about the
taste questions about what things they’d
like to change gathering feedback and
all the person drinking the beer has to
do is follow a link printed on the
bottle and the algorithms start talking
for them so by combining these two sides
product and consumer if you like left
brain and right brain what we have
created are algorithms that learn which
recipes work on which time they learn
which ideas to try in the future on
which to retire
they learn what to make for individuals
and they learn what to make for groups
now I think the real beauty in this
whole thing is that we’ve taken
something inherently artificial and
complex and we’ve brought the whole
process of creation and mass consumption
back to its core human components there
is a direct visceral connection between
the craftsman and the consumer almost as
if they were in the same room now whilst
we want to take and exploit the wisdom
of the crowd it isn’t as simple as
averaging and in fact the key complexity
is being able to interpret what each
individual says unfortunately we’re not
all as wise as each other and so we have
to be careful whose opinion we take and
whose we don’t but that’s where AI
really comes in because it learns which
questions to ask to which people and it
learns how to understand the answers so
when after one of our early beers we the
algorithms ask people whether they like
more or less alcohol in the recipe the
vast majority of people said they wanted
a stronger beer so we made a stronger
beer and what happens people actually
liked it less so the algorithms learn an
important lesson
well they learnt there we’re a nation of
alcoholics of course right but by making
a mistake and seeing what happens they
learnt to better interpret people in the
future the algorithms aren’t perfect
they do make mistakes but by exploring
and gathering feedback they learn not to
make them again they learn by doing they
learn by trying things over and over
again iterating and changing I believe
that it is in creating a constantly
changing product that the solutions are
all these problems lie
without change there are no mistakes for
Rai to learn from no mistakes for us to
learn from without change
we can’t explore the whole universe of
product possibilities we can’t make
things better and without change we are
not selling the same thing again and
again and again and actually all sorts
come down to you we go back to
advertising instead of innovating so the
beer is always changing
we started with one beer which became
four beers and those four beers spawned
a further 18 different age iterations
all of them different I was talking to
some of the other day and I was
explaining that we never make the same
batch twice but if you’d loved one of
these beers then you’re gonna have to go
home and make it yourself and he looked
at me like it was crazy
what if you through all that discovered
that there was a perfect bit you
discovered the perfect beer and the
answer is simple I don’t believe that we
can make a perfect beer that isn’t to
say that we can’t make a perfect beer I
don’t believe it exists and that comes
down to people people want to try new
things they’re obsessed with the idea
that there’s something better out there
that I haven’t yet had above all they
want new experiences so we have to keep
on creating it’s all about people and
really all this is about people I
started working in AI because I was
interested in people I was interested in
the way the human brain works how we
interact with each other I am interested
in AI not because it can replace us
but because it can bring us closer
together human creativity is still very
much at the heart of this process but
the crucial difference now is when the
brewer says I have an idea an AI
responds and says well let’s go and
check it out if it’s a good one and
replace brewing with any type of product
innovation you care to imagine by using
artificial intelligence we can eradicate
the deep inefficiencies of today
this is a road to a future full of new
experiences not just tastes and smells
and whilst a few of us may end up
getting run over by self-driving cars it
will be AI that gets us there the Mad
Men can finally be disrupted because we
have a choice we all know that
artificial intelligence is gonna change
our lives but what is clear now is in
terms of the how and the when of change
it will be up to the people to decide
thank you very much
[Applause]
just launched Barbie the madmen are
revolutionising car advertising and you
can still smoke in your office if you
work at a brand in 1960 and you’re
trying to create the next greatest
product and you’ve got to be able to
predict where your customers want so you
go out and talk to them you stand in a
mall and hand out questionnaires you run
a focus group you do market research
fast-forward 60 years later to today and
companies still need to talk to their
customers and they still do it in
exactly the same way so think of the
things that you bought this week I got
some new clothes pair of headphones a
can of coke almost all the things that
you bought would have been created out
of our discussion with small groups of
people using pretty much the same
techniques as 60 years ago and that
isn’t because they work they frequently
don’t the reality is the vast majority
of new products fail the wrong things
just get built and so we live in a world
that is biased away from innovation
towards promotion last year we spent
five hundred billion dollars on
advertising and that’s money that is
being used to convince us what to buy
but that does next to nothing to
actually improve the product and
companies are constantly spamming us
telling us what we want and it works
you’re actually looking at a man who now
buys more expensive shampoo than he did
when he had a full head of hair
that is a true story so what if the
world was less wasteful what if we could
create a future where all the energy
that we currently spend on advertising
was instead spent on making a better
product a future where companies were so
good at listening to us that their
innovation replace the need for
advertising altogether in that future
marketing isn’t about pushing and
promotion it’s about managing the flow
of information between those that make
products and those that consume products
now I believe that this feature is
possible and the key is our relationship
with artificial intelligence at the very
core of AI we’re trying to create
machines that learn machines that learn
in many ways like you and I do and in
fact one of the main things that these
machines that need to learn about is us
but whilst you and I are pretty limited
on the number of people that we can
learn about at the same time the
machines that we’re designing can
observe and interact with millions of
people simultaneously they can combine
that knowledge and form very accurate
models about how each of us behaves and
there are plenty of these machines
already on the internet constantly
learning about us watching what we click
on what we buy who we connect with and
the longer they watch the better they
get at predicting where we might click
next what we might like to buy who our
friends might be we’re used to this
whole idea of what we’ve now become has
become known as personalization highly
tuned adverse chases around the internet
however the products they are selling
don’t get better and so we’re back to
where we were in 1960 sure we’re far
more connected now but our personal data
is being used in a way that we can’t
readily track to form incredibly
sophisticated models of our purchasing
behavior and those models are then used
to convince us to buy more things and
they still do not improve the products
that which one are creates creates a lot
of value for businesses but nowhere near
enough value for us consumers
so last year we said about trying to
rethink this we wanted to see if in the
future the most effective brands are the
ones that talk to their customers not to
make better advertising but to share
ideas we thought that they’d use
artificial intelligence to help real
people and brands talk to each other and
we wanted to prove this in an industry
which people have very strong views on
and that which we had a pretty
significant interest in as well
beer so we created intelligent X the
world’s first beer brewed by AI by
incorporating AI into the brewing
process we have created a product that
automatically changes its recipe based
on the needs of the consumers a beer
that continually adapts to the changing
tastes of the people drinking it now
that is what I call personalization and
it works very simply on one side
algorithms are deciding which
ingredients to use how long’s to cook
things for and on the other they’re
asking questions for the people actually
drinking the beer questions about the
taste questions about what things they’d
like to change gathering feedback and
all the person drinking the beer has to
do is follow a link printed on the
bottle and the algorithms start talking
for them so by combining these two sides
product and consumer if you like left
brain and right brain what we have
created are algorithms that learn which
recipes work on which time they learn
which ideas to try in the future on
which to retire
they learn what to make for individuals
and they learn what to make for groups
now I think the real beauty in this
whole thing is that we’ve taken
something inherently artificial and
complex and we’ve brought the whole
process of creation and mass consumption
back to its core human components there
is a direct visceral connection between
the craftsman and the consumer almost as
if they were in the same room now whilst
we want to take and exploit the wisdom
of the crowd it isn’t as simple as
averaging and in fact the key complexity
is being able to interpret what each
individual says unfortunately we’re not
all as wise as each other and so we have
to be careful whose opinion we take and
whose we don’t but that’s where AI
really comes in because it learns which
questions to ask to which people and it
learns how to understand the answers so
when after one of our early beers we the
algorithms ask people whether they like
more or less alcohol in the recipe the
vast majority of people said they wanted
a stronger beer so we made a stronger
beer and what happens people actually
liked it less so the algorithms learn an
important lesson
well they learnt there we’re a nation of
alcoholics of course right but by making
a mistake and seeing what happens they
learnt to better interpret people in the
future the algorithms aren’t perfect
they do make mistakes but by exploring
and gathering feedback they learn not to
make them again they learn by doing they
learn by trying things over and over
again iterating and changing I believe
that it is in creating a constantly
changing product that the solutions are
all these problems lie
without change there are no mistakes for
Rai to learn from no mistakes for us to
learn from without change
we can’t explore the whole universe of
product possibilities we can’t make
things better and without change we are
not selling the same thing again and
again and again and actually all sorts
come down to you we go back to
advertising instead of innovating so the
beer is always changing
we started with one beer which became
four beers and those four beers spawned
a further 18 different age iterations
all of them different I was talking to
some of the other day and I was
explaining that we never make the same
batch twice but if you’d loved one of
these beers then you’re gonna have to go
home and make it yourself and he looked
at me like it was crazy
what if you through all that discovered
that there was a perfect bit you
discovered the perfect beer and the
answer is simple I don’t believe that we
can make a perfect beer that isn’t to
say that we can’t make a perfect beer I
don’t believe it exists and that comes
down to people people want to try new
things they’re obsessed with the idea
that there’s something better out there
that I haven’t yet had above all they
want new experiences so we have to keep
on creating it’s all about people and
really all this is about people I
started working in AI because I was
interested in people I was interested in
the way the human brain works how we
interact with each other I am interested
in AI not because it can replace us
but because it can bring us closer
together human creativity is still very
much at the heart of this process but
the crucial difference now is when the
brewer says I have an idea an AI
responds and says well let’s go and
check it out if it’s a good one and
replace brewing with any type of product
innovation you care to imagine by using
artificial intelligence we can eradicate
the deep inefficiencies of today
this is a road to a future full of new
experiences not just tastes and smells
and whilst a few of us may end up
getting run over by self-driving cars it
will be AI that gets us there the Mad
Men can finally be disrupted because we
have a choice we all know that
artificial intelligence is gonna change
our lives but what is clear now is in
terms of the how and the when of change
it will be up to the people to decide
thank you very much
[Applause]
Please follow and like us: