All language subtitles for Secrets.Of.Big.Data.S01E06.WEBRip.x264-ION10

Narrator: A Spanish badminton
player is criticised

for using artificial
intelligence

in an attempt to make history.

Ramona Pringle: She really is
the epitome of a modern athlete.

Anthony Morgan: None of her
rivals have access

to these huge databases
of information.

Is that fair?

Narrator: An American man is
convicted of a crime based on

incorrect results from
a lie detector test.

Mhairi Aitken: He was obviously
nervous about his arrest

and was distraught at the time.

Narrator: But using Artificial
Intelligence systems

may prevent this from happening
in the future.

Nikolas Badminton: They have
a data set to draw from

to determine if a suspect is
being truthful--

their eyes, voice and even
their brain activity.

Aitken: In theory,
if this worked,

criminals would be
more easily convicted

and innocent people
wouldn't be sent to jail.

Narrator: For fun, an Indonesian
student offers his selfies

as digital art for sale

as non-fungible tokens.

Morgan: Some prominent
crypto influencers

started promoting
Ghozali's project

and the prices skyrocketed.

Pringle: His little joke project
has him laughing

all the way to the bank.

Narrator: These are the stories
of the future that big data

is bringing to our doorsteps.

The real world impact of
predictions and surveillance.

The power of artificial
intelligence

and autonomous machines.

For better or for worse,

these are the
Secrets of Big Data.

♪ [show theme music]

♪♪

Spanish badminton sensation
Carolina Marin has one goal--

to become the first
person ever to win

the European Championships
five times in a row.

On May 2nd, 2021,

one win away from
achieving thi sgoal,

she takes to the court
at the Palace of Sports

in Kyiv, Ukraine.

Morgan: This woman
has got skill.

By this point, she had already
won three world championships

and a gold medal at
the 2016 Olympics.

Narrator: Marin is talented
and tenacious,

practicing for countless hours

in order to reach the pinnacle
of her sport.

But she also has a big
weapon in her arsenal--

Big Data.

The Spaniard uses technology in
all aspects of her preparation.

From analysing statistical
datasets to fine tune

her training and monitor her
physical fitness,

to predictive algorithms that
generate strategies

to use against her opponents.

Pringle: She really is the
epitome of a modern athlete.

One that has fully
embraced the leg up

that technology can provide.

Narrator: As Carolina Marin
attempts

to make badminton history,

the power of artificial
intelligence and big data

in the sporting world is about
to be put to the test.

Badminton: The use of
data and AI in sports

has exploded in recent years.

It's such a data and
stats-driven environment

that it was pretty much
inevitable.

Narrator: Major sports
teams across the world

now have entire departments
devoted to analytics,

which uses statistical data and
mathematical principles

to assess player performance,

business operations,
and recruitment.

Front offices that were once
staffed with former players

and coaches are quickly becoming
the domain of non-athletes,

who are more comfortable sitting
in front of a computer

than on the field.

Morgan: Some people don't
like these developments,

they say these analytics don't
provide a full enough picture

of an athlete's potential
or performance.

Pringle: In a nutshell,
the jocks are afraid

that the nerds are taking over.

Narrator: Statistical analysis
in sports goes back

over 150 years,

starting in the United States.

In the late 1850s, Henry
Chadwick, a sportswriter,

developed a metric for baseball
called the box score,

which he adapted from a
cricket scorecard.

The box score showed a player's
results in table form,

making it easy to
measure performance

based on statistical data.

Morgan: Chadwick invented
metrics that are still used

to this day--earned run
averages for pitchers

and batting averages
for, well batters.

Badminton: He was the first
person to recognize

that mathematical analysis
could be applied

to accurately evaluate
a player's abilities.

Narrator: A century
later in England,

individual player performance
starts to give way

to a new sphere of analytics--

team tactics.

In 1950, Charles Reep,
an accountant

and member of the
Royal Air Force

decided to apply
his maths acumen to soccer.

Pringle: He was watching a game
and became frustrated

by the endless attempted attacks
that didn't result in goals.

So he did what any good
accountant would do--

he grabbed a pencil and
started making calculations.

Narrator: Reep counted
the number of shots

and passes in games
and then used that data

to conclude that the
majority of goals

were scored after a build
up of four passes or less.

Although his theory
was later questioned,

Reep's research became the
foundation

for the long-ball approach--
with the goalkeeper

or defender playing the ball
directly to an attacker--

that dominated English soccer
in the years that followed.

Morgan: It is an important
turning point

in sports analytics -

understanding this data
actually changed

the way that this
game was played.

Narrator: As technology advanced
into the computer age,

huge databases of
historical statistics

across all sports
became available.

Badminton: Every game that's
ever been played

in any sport leaves
a data trail.

Now multiply that
with the number of games

that have been played
over the last 75 years

and the amount of
information is remarkable.

Narrator: Building on the
pioneering work of Charles Reep,

the analysis of this mountain of
data has led to conventional,

time honoured strategies
being abandoned

because they don't
pass the data test.

In 2007, Daryl Morey was hired
as General Manager

of the Houston Rockets of the
National Basketball Association

or NBA, despite never having
played or coached in the league.

But what he did have was a
computer science degree

and a Master's in Business
Administration.

Pringle: Neither of
which mean anything

to basketball's old guard.

Badminton: After a deep dive
into the data, he concluded

that it's more efficient
for teams to attempt

three point shots
instead of two pointers.

It seems kind of obvious,

the more three
pointers a team takes,

the more potential points
they will be able to score,

and the more games
that they'll win.

Narrator: This philosophy
resulted in a fundamental shift

in basketball strategy, with
three point shooting wizards

like Steph Curry becoming the
preeminent stars of the NBA.

♪

And in the National
Football League,

data analysts who had long been
pushing for teams to stop

giving up possession by punting
the ball in certain situations,

are finally being heard.

Morgan: In the 2020 NFL season,
teams punted an average

of just 3.7 times per game,
which is the fewest number

of times since we started
recording statistics on punting.

After studying the data, people
figured out that punting

is kinda of a lousy strategy,
especially if it's late

in the game and your team
is already behind.

Narrator: But critics warn that
there are dangers in giving

too much credence to analytics,
particularly when it comes

to evaluating young athletes
who are still developing.

Badminton: Data doesn't
tell the whole story.

It can't recognize a player's
potential for improvement,

or their dedication
and willingness

to go the extra mile
to get better.

These things can't easily be
measured using statistics.

Pringle: It's become too easy
for teams to give up on someone

if the numbers tell them to,

but the sports world is
full of late bloomers

who may never have gotten a fair
shake in the analytics age.

Narrator: Others would
argue that the opposite

can also be true, that
some players benefit

from underlying data
that clearly indicates

they are still a valuable asset
to the team,

even if they don't appear
in highlight reels,

or on the front page
of the sports section.

Pringle: These are what's
called "analytics darlings".

Their value may not be
evident to the average fan

because they operate
outside of the spotlight,

but the data proves their
importance to the team.

Narrator: However, not all
analytics is about evaluating

player performance
and changing tactics.

Sports are a huge
business after all,

with billions of
dollars at stake,

so naturally teams want to use
technology to maximise profits.

Morgan: This is
especially important

when negotiating contracts.
Because now,

teams can evaluate their
players against others

to figure out what a fair
market price really is.

Narrator: And sometimes the
shoe is on the other foot,

with tech-savvy players
turning to big data

to ensure that they don't get
short-changed

by exploitative management.

In 2021, Manchester City star
Kevin De Bruyne

employed an analytics company
rather than an agent,

when negotiating a new contract.

Badminton: The company's
software analysed his past,

present and projected
future performances

and then compared them to
other world-class midfielders.

The data determined that
De Bruyne was underpaid

by a wide margin.

Morgan: He ended up signing a
contract making him

the Premier League's
highest paid player.

One - nil, analytics.

Narrator: De Bruyne was
one of the first players

to use technology
in this manner,

he was an early adopter,
much like Carolina Marin.

Pringle: Marin and her team were
using data analytics

as far back as 2006,

when she was only 14 years old.

They studied opponents'
game footage and entered

the information manually
into an excel spreadsheet,

learning her rivals' patterns so
that she had an advantage

before even stepping
onto the court.

Morgan: Obviously this can be
done with predictive algorithms,

which means that the process is
not only faster,

it's more comprehensive.

But the result is the same.

Marin goes into each game
with a sound strategy

based on historical data.

Narrator: The AI translates
Marín's and her opponets

movements into data, designed
to enhance performance,

forecast the pattern of a game

and exploit her
adversaries' weaknesses.

Badminton: The algorithm
may determine

that her opponent
loses most points

that are played deep
to their backhand side.

Or maybe they lack foot speed
and drop shots are effective.

It can also tell
her when to attack

and when to play
more defensively.

Pringle: On the predictive side,
the system is able to foresee

what the opposition is most
likely to do

in specific situations.
For example,

they may serve in different ways
whether they are winning

or losing at certain
points in the match,

and Marin is able to anticipate
and respond accordingly.

Narrator: And as the final
match in Kyiv progresses,

Marin's preparation is making
things very difficult

for her young Danish opponent.

The Spaniard cruises
to a 21-13 victory

in the first set of the best
two out of three match.

Her record-setting fifth
straight European Championship

is now well within reach.

But not everyone is
cheering for Marin.

Her use of technology is not
without its detractors,

with some critics believing
that the playing field

is not always level.

Morgan: None of her rivals
have access

to these huge databases
of information

and you could ask,
'Is that fair?'

I mean, some of these
other teams might not have

the technology or even
be able to afford it.

Badminton: But the data is out
there to be collected

and AI systems are
becoming cheaper

and more readily available to
those that want to use them.

This is the future
of the sport and those

that don't take advantage of it
will get left behind.

Narrator: And there are
broader issues to consider.

Some believe that the use of
technology in sports

contradicts everything
that fans embrace--

spontaneity, creativity
and athletic instinct.

Pringle: When someone like Marin
steps on the court knowing

what her opponent is going to do
because a machine told her,

to some it cheapens her success.

Narrator: But those who defend
the use of AI in sports,

argue that sports aren't
immune to change,

and should evolve along with our
relationship to technology.

Morgan: Look at equipment.

No tennis players use wooden
racquets any more,

but the first person
to use the metal one

would have been way
ahead of the curve.

Maybe we're just looking
at the same thing here.

Narrator: If there's one area of
AI use in athletics

that nobody can take issue with,

it's in the prevention and
treatment of injuries.

The advanced predictive and
diagnostic capabilities

of these systems make them
ideally suited

to protect the health of their
most valuable assets.

Badminton: Many teams now use
wearable devices

to track players' movements
and physical exertion.

The AI systems then analyse
the collected data stream

to identify warning signs
of potential injury.

Narrator: In the
2015/2016 season,

underdogs Leicester City
shocked the sports world

by winning the English
Premier League,

thanks in part
to the use of data.

Players wore a small device in
training during matches,

that collected information on
things like distance covered,

acceleration, heart rate,

and impact of collisions
with other players.

Pringle: It helped staff monitor
the wear and tear on players,

and they had fewer injuries than
any other team in the league.

It's certainly not the only
reason for their success,

but it was definitely
a contributing factor.

Narrator: A little more than 150
kilometres away in Liverpool,

one of Leicester's
Premier League rivals

recently teamed up with a
company called DeepMind

in order to experiment with the
use of artificial intelligence.

Morgan: Liverpool supplied
DeepMind with data

from every one of its matches
played

between 2017 and 2019,

hoping that the AI
would find patterns

that would give them
advantage in competition.

Badminton: There's a lot of
uncertainty in team sports,

which presents a challenge
for researchers.

A game like chess has
a basic set of rules

so there are a finite
number of outcomes,

but there are so many different
permutations in a soccer match

that it's difficult to predict.

Narrator: Difficult,
but not impossible.

DeepMind's research established
that it's possible to use data

to teach an AI system to predict
how individual players

will react in response
to certain actions.

Morgan: Who knows what
the AI might find?

It might tell you exactly what
a player in the midfield

with the ball would do
before they even know.

Pringle: This means that you
can develop strategies

that are tailor-made to use,

not just against specific teams,

but based on which
individual players

are on the field
at any given time.

Narrator: DeepMind's study also
analysed

over 12,000 penalty kicks

taken around Europe
over the last few years

and then used the data
to forecast the odds

of taking a successful penalty,

based on where players were most
likely to place their shots.

Morgan: They found that strikers
are far more likely to go

for the bottom left-hand
corner of the goal

as opposed to midfielders
who often mix things up.

Badminton: This kind of
information is valuable

to goalkeepers because they must
get where the shot is going,

to have a chance at
making the save.

If they have the relevant
data going into the match,

they may be able to guess
correctly more often.

Narrator: And back in Kyiv,
Carolina Marin seems to be

guessing correctly with
astounding frequency.

Marin overpowers her opponent

and takes the second set 21-18,

capturing the title.

It is a victory
for the record books,

and one that establishes
her as arguably

the greatest player
of her generation.

Pringle:
Marin's achievement goes to show

that the use of technology
can have a significant

positive impact in sports
at the highest levels.

Narrator: But it isn't just
at the highest levels

that AI can make a
difference in athletics.

As the technology becomes
cheaper and more widespread,

weekend warriors and
aspiring pros

are starting to join the
technology revolution.

In the first two years after its
launch in 2018,

HomeCourt, a free iPhone app
that analyses

basketball players as they
practice shooting,

tracked over 60 million shots

from over 170 countries
around the globe.

Morgan: It's a pretty
simple system, really.

Just point your camera at
yourself while practicing

and the app tracks the location
and success rate of your shot.

Then it provides you feedback
in real-time.

Pringle: But it does
have limitations.

It has trouble
performing if there's

more than one person on
the court, and even though

the company has taken steps
to address that issue,

the AI at the app level
just isn't quite there yet.

Narrator: Even though Home
Court's program has some flaws,

it symbolises a critical
step forward

in access to technology
for athletes.

In the not so distant future,

the need to spend exorbitant
amounts of money

on elite coaching and assistive
technology

in order to compete
at a high level,

may be eliminated or reduced,

thanks to AI-driven apps.

Badminton: I think what people
want is a level playing field,

an overall democratisation
of technology use

within the athletic community.

And every day, we're inching
closer towards that goal.

Narrator: On the heels of her
electrifying victory in Kyiv,

Carolina Marin's dream of
defending her Olympic title

at the 2021
Olympic Games in Tokyo

was dashed due to a knee injury
suffered in training.

But there's little doubt
that technology will play

a significant role in
her rehabilitation.

♪ [show theme music]

♪♪

Narrator: In Elizabeth,
New Jersey,

22-year-old Emmanuel Mervilus,

and his friend, Daniel Desire,

have just left a donut shop near
the train station.

The two men are walking down
the street when suddenly,

two police officers stop them.

It turns out a 26-year-old man,
Miguel Abreu,

was just robbed and
stabbed nearby,

and the two friends
are the prime suspects.

[police sirens]

Aitken: Abreu says
Emmanuel held him,

while Desire stabbed him
once in the chest

and then they fled
with his backpack.

Narrator: Mervilus and Desire
deny the allegations,

but Abreu is certain
it was them.

They are subsequently arrested
and Mervilus is charged

with armed robbery and
aggravated assault.

His friend is charged
with theft.

Kris Alexander: Mervilus is
confident that they'll get off,

because he says they
were misidentified.

[slamming gate]

♪♪

Narrator: To prove his
innocence,

Mervilus volunteers
to take a polygraph test.

Aitken: Polygraphs use sensors
to measure a person's breathing,

pulse, blood pressure and
perspiration

when responding to questions.

Narrator: To his shock,
Emmanuel Mervilus

fails the polygraph test

and when he asks to take it
again, his request is denied.

Alexander: Law enforcement
experts say Polygraphs are

rarely ever wrong and are often
cited in courtrooms as evidence.

Narrator: Some legal experts
claim Polygraph machines are

accurate in detecting deception
80 to 90% of the time.

And at Emmanuel's trial,

the officer who
administered the test

tells the jury that
the results indicate

that he wasn't telling
the truth.

Mervilus is convicted of
first-degree robbery

and aggravated assault

and sentenced to 11 years
in prison.

His friend, Daniel Desire
is sentenced to four years.

Aitken: Mervilus continually
denies

robbing and assaulting Abreu

and is adamant that
the polygraph was wrong.

Narrator: More than
2.5 million Polygraph tests

are conducted each year
in the US.

And they are also widely used
in Canada, India,

Australia and countries
across Europe.

Mervilus' lawyer echoes his
client's sentiments

and maintains that the lie
detector test was inaccurate.

And as it turns out,
he may be right.

During the trial, the victim is
unable to identify Mervilus

as the person who assaulted
and robbed him.

Aitken: When he was asked to
pick out his attacker,

Abreu pointed to another man who
was sitting in the gallery.

Narrator: In light
of this evidence,

Mervilus' lawyer files
a motion for a new trial.

It's quite possible that
the polygraph results

were misleading, calling into
question the

overall accuracy of
lie-detection machines.

But what if artificial
intelligence based systems

could ensure that something like
this never happens again?

Badminton: AI could be far
more reliable in determining

if a suspect is either lying
or telling the truth

than a traditional
polygraph machine.

Narrator: AI lie detection
systems use machine learning

to analyze a suspect's
blood pressure,

pulse, breathing rate,

as well as their body movements,
and facial expressions,

all of which may indicate
that they are lying.

Alexander: It's also able to
extract information from

audio frequencies in a
suspect's vocal patterns.

Narrator: There are a
significant number of

researchers and private
companies that are developing

AI powered 'Lie Detectors'.

Researchers at Cornell
University in New York

claim they have developed a
system

that uses courtroom trial videos
to detect deception

with an 87.7% accuracy rate.

And recently a U.S. company
developed a system

that focuses on the organ that
will most likely reveal

when someone's lying--the eyes.

Badminton: With a precision
optical scanner,

algorithms can evaluate minute
changes in eye direction

and pupil diameter when people
respond to questions.

Narrator: One such program,
'Eyedetect',

is being used by more than
65 US law enforcement agencies

and nearly 100 others worldwide.

The companies behind such
AI lie-detection tools

claim they will make our world's
legal systems

more efficient and fair.

Aitken: It could have
a significant impact.

In theory, if this worked,

criminals would be
more easily convicted,

and innocent people
wouldn't be sent to jail.

Narrator: But many critics
point out that there's

no empirical evidence suggesting

that there are universal
physiological

or physical reactions when a
person's being deceptive.

Alexander: An innocent person
may be nervous

when taking the test;
their heart is racing,

they're sweating,
and because of that,

the AI may interpret
this as being deceptive.

Narrator: Emmanuel Mervilus
claims this could be

the reason why he failed
his Polygraph test.

Aitken: While it was being
administered,

he was obviously nervous about
his arrest and he was

distraught at the time because
his mother had recently died.

Narrator: After spending more
than 3 years in jail,

Mervilus finally receives
some good news.

The Superior Court
overturns his conviction

and orders a new trial.

Alexander: He rejects the
prosecution's offer

to plead guilty in exchange
for time already served.

As far as he's
concerned, he's innocent.

But it's still possible
he might not be.

Badminton: It's well-known
that humans do possess

a tremendous capacity
to believe their own lies.

Narrator: There are countless
examples of people

who were actually able to
intentionally deceive

lie detection systems.

Probably the most famous case

was that of serial killer
Ted Bundy.

Aitken: He passed
a lie detector test

despite knowing he had
killed many women.

Narrator: How criminals are
sometimes able to

outsmart these machines,
is actually very simple--

they do it by manipulating their
physical responses

when answering questions
people are likely to lie about,

such as
"Have you ever stolen money?" or

"Have you ever
cheated on a test?"

Alexander: There have been
cases where a subject will

do something like
bite their tongue hard,

or step on a tack in their shoe

when answering dishonestly
to questions that

aren't related to the crime
they're being investigated for.

Narrator: The pain increases
blood pressure,

heart rate, and breathing,

so they hope their
physical response

to non-related questions

will be more elevated
than when they lie,

meaning their lies will be
interpreted as the truth.

Aitken: Another way to beat
a lie detector is to

force yourself to think of
something cheerful while lying.

This will help mute your
physiological responses

and disguise a lie
from the truth.

Narrator: But technology experts
claim that lie detectors

powered by AI won't be fooled
by these tactics.

Badminton: Unlike polygraph
machines,

they have a data set
to draw from

to determine if a suspect
is being truthful--

their eyes, voice,
and even their brain activity.

Narrator: And AI's unique
ability to detect deception

is not only being used
by law enforcement.

Banks, Welfare Offices and
Insurance Companies

have also adopted the
technology.

Alexander: Some Insurance
companies are now requiring

their customers to make
claims with a video

so algorithms can analyze it
and flag specific physical cues,

which may suggest
they're being fraudulent.

Narrator: Currently,
over half of the

top US insurance companies

and 8 of the top 10 banks

are using this technology
to detect fraud.

But Human rights activists are
concerned that as these

AI lie detection programs
become more pervasive,

there may be dangers they could
unleash on our society.

Badminton: One aspect
critics point to,

is there may be inherent bias

in how algorithms
determine deception.

Narrator: Recent research
suggests that a majority

of the subject data used to
train various systems

was gleaned from Caucasian men.

Aitken: They were white
Europeans and Americans,

so it's completely unbalanced
in gender and ethnicity.

Narrator: This could mean that
the AI might inadvertently

discriminate against
people on the basis

of gender or ethnic origin.

Alexander: After all,
it's human beings

who are writing
these algorithms,

and many people do have 'bias'

whether they're
conscious of it or not.

Narrator: The question of
whether or not these

AI lie detection systems
are inherently biased

may never be answered.

Badminton: Because of
copyright considerations,

many companies that
produce these programs,

have not been transparent in how
they determine deception.

But that may soon change
due to legal challenges.

Narrator: In 2021, the European
Court of Justice was petitioned

to obtain previously
undisclosed information

on a controversial new AI
lie detection program.

Aitken: It's an EU funded
research project,

which is being developed to
assess the trustworthiness

of travellers and
immigrants at borders.

Narrator: The application is
called 'Avatar'

and it acts as a
virtual border agent.

When travellers go
through customs,

it asks them a variety of
questions

and using a motion sensor and
infra-red eye-tracking camera,

measures how they respond.

Alexander: It's apparently
able to assess

if a traveller is lying
within 45 seconds.

It's being touted as the
future of passport control.

Narrator: But even though
programmers deny this,

there are worries the system
may be discriminatory.

Furthermore, some programmers
claim the benefit the technology

has in accurately
detecting deception

far outweighs any potential
problems it may cause.

Badminton:
Sometime in our near-future,

these systems may be able
to achieve 100% accuracy.

Narrator: That may be good news
for people who've been

falsely convicted of a crime -
people like Emmanuel Mervilus.

After reviewing the evidence
at his retrial,

the jury acquits him
of all charges.

Alexander: In light of his
wrongful conviction,

Mervilus is eventually
compensated

close to $100,000 by
the state of New Jersey.

Aitken: The hope is that AI can
prevent this from happening.

The goal is to accurately
delineate 'truth' from 'fiction'

so no innocent person
ever goes to jail again.

Narrator: That may
sound good in theory,

but leading academics and
scientists wonder,

what will our world look like

if no one is able to lie
anymore?

Badminton: The truth is, not
all lying is actually bad.

Many of us do it to
protect people's feelings,

like telling children
"everything is fine"

to preserve their
sense of security.

Narrator: Many fear that if
this ability is taken away,

it will have a profound
effect on humanity,

especially if a reliable lie
detector is made available

to the general public,
via a smartphone app.

Aitken: It would completely
change how we interact

with each other on
a day-to-day basis.

No one would get away
with anything anymore.

Narrator: Whether or not AI
powered lie detection systems

can ever achieve true precision,

is a highly contentious subject.

Some say that lying is just too
complicated and individualised

for any machine to
conclusively recognize,

regardless of how sophisticated
the technology is.

Others counter that 100%
accuracy in lie detection

is not only possible,
but imminent.

Which side is telling the
truth remains to be seen.

♪

♪ [show theme music]

♪♪

Narrator: 22-year-old Indonesian
Computer Science student

Sultan Gustaf Al Ghozali
has a daily ritual.

For the last five years, he has
sat down at his desk

and snapped a selfie using his
computer's camera.

The pictures are all
largely the same,

Ghozali gazing into the lens

with no discernible expression
on his face.

Morgan: He wanted to use the
pictures to make a video

for his graduation.

It was supposed to be a time-
lapse to mark the occasion.

Narrator: Although he has missed
days here and there,

by the end of 2021,

Ghozali has amassed a stockpile

of nearly a thousand photos.

Pringle: As something of a joke,
he decides to see

if he can sell his selfies

as Non Fungible Tokens
on OpenSea,

the largest internet exchange
for buying and selling NFTs.

Narrator: NFT's are
taking the digital art

and collectables world by storm.

♪

Alexander: They're like modern
day collector's items,

similar to old paintings or
vintage sports cards.

The only difference is that
you're buying a digital file

with verification that you are
the owner of the original work

instead of a physical object.

Narrator: Each NFT can only have
one official owner,

so to prove its originality,
every single NFT has a

singular unit of data called a
smart contract attached to it.

This smart contract certifies
that the asset is authentic,

impossible to duplicate,

and easily traceable.

Morgan: Traditional certificates
of authenticity or bills of sale

are pretty easy to manipulate.

But NFTs have code
built right into them

so it makes that impossible.

Narrator: NFTs are registered on
a blockchain called Ethereum,

a digital ledger which records
information in a way

that makes it extremely
difficult to change,

hack, or cheat the system.

The NFT's themselves are bought
and sold using

Ethereum's cryptocurrency,
Ether.

Thinking that nobody would have
any interest in buying them,

Ghozali lists his photos
for sale at a mere

one hundred-thousandth
of an Ether,

equal to about $3 at the time.

But over the next several weeks,

his collection would
inexplicably go viral

and come to exemplify the weird

and wonderful world of NFTs.

♪♪

In 2014, the first
acknowledged NFT

was created by digital artist
Kevin McCoy.

The work, "Quantum",

is a pixelated image
of an octagon,

filled with different shapes
hypnotically pulsing

in bright colours.

The work later sold at Sotheby's

for $1.47 million US dollars

in June of 2021.

Alexander: After the creation
of Quantum,

the next several years
saw a rapid expansion

in the popularity of NFTs.

The underlying technology
got much better,

and a new market
for art was created.

Narrator: In March of 2021,
Mike Winkelmann,

the digital artist
known as Beeple,

stunned the art world when
one of his NFTs sold for

an astounding
$69 million dollars,

through the first ever NFT sale
at Christie's Auction House.

Morgan: Six months earlier,
the most Winkelmann

had ever sold a print for
was about $100 bucks.

But now, he was among
the most profitable

living artists in the world.

Narrator: The work called

"Everydays:
The First 5,000 Days",

is a collage of
Winkelmann's output

from the previous 13 years,

when he produced and published a
digital artwork

every day for a project he
called "Everydays."

Ghozali, in a subtle nod to
Winkelmann,

calls his collection
"Ghozali Everyday"

when he lists his selfies
for sale

on the OpenSea platform.

Morgan: Not much happened with
Ghozali's project

in the first few days, until

a celebrity chef from Indonesia

who also happened to be an NFT
enthusiast stumbled upon it.

Narrator: Arnold Poernomo,

best known as a judge on
MasterChef Indonesia,

tweets about Ghozali's
collection

and other NFT influencers
quickly follow suit.

Interest in the project suddenly
starts to gain momentum.

Pringle: Ghozali
is kind of shocked

when 35 of his selfies
sell in one day.

Little did he know,
it was just the beginning.

Morgan: The thing
about NFTs is,

demand comes out of nowhere.

Narrator: Critics claim
that in some cases,

this demand is produced
fraudulently,

through a process called
wash trading.

Alexander: Wash trading is when
a person is on both sides

of a transaction, acting as
both buyer and seller.

This creates the illusion of
demand and high trading volume,

which drives the value
of an asset up.

Narrator: Trading in NFTs has
skyrocketed,

rising from $106 million
dollars in 2020

to an extraordinary
$44.2 billion in 2021.

And although 'wash trading'
is a growing concern,

most NFT supporters are
quick to downplay it.

Pringle: One study found that
over half of the people

who engaged in wash trading,

actually lost money due to
what's called "gas fees",

which are paid by traders
to offset the energy required

to process transactions on
the Ethereum blockchain.

Narrator: Energy consumption is
becoming a significant problem

within the NFT community.

The Ethereum blockchain handles

over 1 million transactions
per day,

and it has been estimated that
the average transaction

uses 35 kwh of power,

roughly the same as running a
refrigerator for a month.

Morgan: And that's just
for the transaction.

If you add up the total process

for creating and transferring
these NFTs,

it can be ten times that amount.

Alexander: One particular
example showed that an artist

who sold two works used over
175 Mwh of energy, equal to

21 years of emissions for the
average American household.

Narrator: By current estimates,
Ethereum is thought to consume

around 112 terawatt hours of
energy annually,

which is more energy
than countries

like the Philippines or
Pakistan consume in a year.

The resulting carbon emissions
are also troubling,

around 62 metric tons,

roughly the same footprint
as Belarus.

Pringle: It's a really big
problem, but steps

are being taken to address it.

NFT supporters argue that
the market's popularity

will incentivize
the shift of energy grids

from traditional fossil
fuels to renewable sources.

Morgan: It's an understandably
confusing argument.

I mean it sounds like they are
saying: we wanna save the planet

and to do it, we have
to keep damaging it.

Narrator: Critics also point out
that the problems with

the NFT market go beyond
environmental damage,

claiming that in its
current iteration,

it's little more than a
high tech pyramid scheme.

Alexander: The market relies on
early investors making money

off of buyers who get in late.

The new money is drawn in by

misleading promises
of big returns.

Pringle: But NFTs have
intrinsic value,

ownership of a digital asset.

Pyramid schemes have
no value whatsoever,

so it's not really a
legitimate comparison.

Narrator: Theft has also
become a major issue,

as unsuspecting artists have had
their digital artwork uploaded

as NFTs to online marketplaces
without their knowledge.

In order to claim ownership,

the thieves simply add a
digital token to the NFTs,

and because the blockchain is
anonymous,

it's virtually impossible for
an artist to get restitution

if their work is stolen.

Morgan: One Dutch artist found
more than 100 pieces of her art

up for sale on OpenSea,

which surprised her because
she didn't post them.

Alexander: For victims of theft,
trying to report stolen work

is time consuming
and complicated,

and exchange platforms have
little reason to take action

because they get a cut
off of each transaction,

regardless
of who owns the work.

Narrator: DeviantArt, an online
community for digital artists

that contains around
500 million pieces,

recently began watching the
blockchain

for copies of their users' work.

It has since sent
90,000 warnings

about potential theft to
thousands of their clients.

Pringle: Scammers are now using
bots to scrape online galleries

and sometimes they steal the
entire collection of an artist

and then post them for sale.

Narrator: But some artists have
spoken out in favour of NFTs,

claiming that the system has
made it both more lucrative

and much easier for them
to sell their work.

Thanks to this new technology,

people who once toiled in
obscurity, living hand-to-mouth

are now raking in
thousands of dollars.

Alexander: NFTs have
created a market

for creative work that
didn't exist before,

and while some of the prices
are definitely inflated,

the bottom line is
that things are worth

what people are willing
to pay for them.

Narrator: And as Ghozali is
quickly finding out,

NFT enthusiasts are willing to
pay big money

for just about anything
if the market dictates it.

Morgan: Some prominent crypto
influencers

started promoting
Ghozali's project

and the prices
skyrocketed.

They reached about 4 Ether,

which is about $12,500.

He sold 230 of these
things in a day.

Narrator: By some reports,
the entire

Ghozali Everyday collection
is worth

an astonishing 374 Ether

or $1.2 million.

Ghozali has also earned over
$100,000

in secondary market royalties

from the buying and selling
of his NFTs.

Pringle: His little
joke project has him

laughing all the way
to the bank.

Narrator: Stories like Ghozali's
are becoming more and more

commonplace in the NFT space.

The exploding market has seen
its fair share of

overnight successes, with people
with little or no experience

in the art world making
large sums of money.

In London in 2021,
12-year-old Benyamin Ahmed

decided to put his coding
skills to good use

and launched an NFT collection
called Weird Whales.

Alexander: He created over 3,000
images of pixelated whales,

each with distinct traits.

Amazingly, the entire collection
sold out in nine hours

and he made over 80 Ether,

at the time,
equal to over $250,000.

Morgan: He also earned 30 Ether,
or about $95,000

from royalties of
secondary sales.

This guy was making bank.

Narrator: Ahmed said he took
inspiration from the iconic

CryptoPunks project, one of the
first collections

to capture the imagination of
the NFT community.

Although all these images
can be duplicated,

only the originals have
any actual value.

Pringle: Owning a CryptoPunk
is a sort of status symbol.

People use them as avatars on
their social media accounts,

and is kind of like owning an
expensive watch

and wearing your sleeves rolled
up so everyone can see it.

Narrator: According to most
experts, CryptoPunks

might just be the most important
NFT project in existence.

And for those involved, it's as
much about the community

that it spawned, as it is about
the collection itself.

Alexander: We're seeing these
projects turn into

little subcultures almost.

They have in-person meet-ups
all over the world,

and there's a sense of being a
part of an exclusive community.

Narrator: And that community is
now spreading

into the celebrity world.

The Bored Ape Yacht Club,
a collection that surpassed

$1 billion in sales
in early in 2022,

has attracted a clientele that
includes the likes of Madonna,

Tom Brady and Eminem.

Pringle: Celebrities feel a need
to be in on the hot new thing,

despite the fact that these
bored apes

seem rather easy to steal.

Narrator: With 10,000
Bored Apes created,

in early 2022, a holder lost

$567,000 worth of NFTs,

in what's called
a direct swap scam.

The perpetrator tricked the
owner into exchanging

expensive 'Bored Ape' NFT's

for what were fake PNG files,

using a third party
trading service.

Morgan: This is what happens
when you're dealing with

digital assets
instead of artwork

that you can physically
hang on your walls.

Stealing your work becomes
as easy as hitting CTRL-C.

Narrator: While the NFT market
is not without its issues,

it still maintains an almost
cult-like devotion

among a younger generation
of collectors.

Morgan: Some people believe
that in the future,

every artwork,
physical and digital,

will have an NFT attached.

Alexander: And some people
believe that paying

hundreds of thousands of
dollars for an illustration

of a goofy looking ape
that can be easily copied

is the definition of insanity.

Narrator: For his part, Ghozali
plans to invest his NFT windfall

and dreams of one day opening
his own animation studio.

And he intends to pay taxes
for the first time in his life,

because according to him, he is
'a good Indonesian citizen.'

♪♪

♪ [show theme music]

♪♪

Narrator: In Omaha, Nebraska,
Christina Hunger works

as a speech pathologist
for children.

Hunger is a life-long dog lover,

and after much careful
deliberation,

she finally decides
to adopt a puppy.

Morgan: She names her 'Stella'

and she is a Blue
Heeler/Catahoula mix.

She's pretty adorable and so
Christina sets out to train her.

Narrator: One day, Christina
notices something significant.

When "Stella' wants to
communicate something to her,

she behaves similarly to the
children she works with.

Pringle: Stella cries
to get attention

and makes gestures when
she wants something,

like sitting near the
door to go outside

or looking at her bowl
when she wants food or water.

Narrator: In her work, Christina
often uses a unique device

to help children communicate.

It uses several buttons that
playback pre-recorded words

when they are pressed.

Aitken: It occurs to Hunger

that she might be able
to teach Stella

to communicate in a similar way.

Narrator: Christina makes
it her mission

to create a communication
device

specifically for her
3 month-year-old puppy.

She purchases several
"recordable buttons"

and attaches them to a
rudimentary board.

The first word she is going to
teach Stella' is "Outside".

Morgan: She records the
word on one of the buttons

and each time she is about to
take Stella for a walk,

she stops and presses
that button several times.

Christina/Device: Come..

- "Outside"
- Outside.

Morgan: And while on the walk
she says the word "outside"

several times so that
Stella can link

that word to that activity.

Narrator: Incredibly, after only
a month of training,

Stella learns to press the
button with her paw

when she wants to go outside.

And over time is able to use the
device to communicate

other words, like "eat,"

"come," "no," and "bye".

Pringle: When Stella wants
something, she simply walks over

to the device and presses
her paw down on a button.

Narrator: Christina shares
videos of

'Stella's achievements'
on Instagram

and before long, she becomes a
social media superstar,

garnering hundreds of
thousands of followers.

Eventually, Stella's incredible
feats draw the attention

of zoologists and experts
in-the-field

of animal communication.

Many are skeptical that Stella
is really speaking to her owner.

Aitken: They say most animals,
especially dogs,

don't have any real
comprehension of 'language'.

So Stella shouldn't
be able to communicate

as effectively as she seems to
be doing in the videos.

Narrator: Some technology
experts say it's possible

and believe that
in the near future,

Artificial Intelligence may be
able to finally close

the communication gap between
animals and humans.

Morgan: It would be a
monumental achievement,

something that the scientific
community

has been working
towards for decades.

Narrator: In an experiment with
dolphins in the 1960s,

the US Navy, along with
independent researchers,

made the first recorded attempts
by humans

to communicate with animals.

Using the limited computer
technology that was

available to them, the team
created an audio communication

interface that was highly
advanced for its time.

Pringle: The system was
programmed to match dolphin

whistle audio waves
with human vowels.

So it can translate the
sequences of vowels

to generate dolphin sounds.

Aitken: Eventually, the dolphins
learned to respond to more

than 30 commands given to them
using five-word sentences.

They also figured out
how to mimic

computer-generated whistles
when prompted.

Narrator: But even though the
dolphins were able to comprehend

basic commands
and imitate whistles,

there was no tangible evidence
that they were actually

communicating with the team
in a human sense.

Morgan: The research showed that
some animals might be capable

of understanding human language,

but only in limited ways.

Pringle: It's an
interesting first step,

although it's still a far cry
from actual communication.

We still don't know if animals
have the capacity for that.

Aitken: But, that still doesn't
mean animals aren't able

to communicate vocally.

Virtually all species
are able to do that,

using a variety
of different sounds.

Narrator: Dogs will growl or
bark to warn others of danger,

a bird will whistle to
attract a potential mate.

And AI may reveal that some of
these animal sounds

may be much more complex than
previously thought.

In March 2020, a group
of scientists

used machine learning algorithms
in an attempt to decipher

the unique sounds produced by
sperm whales.

Morgan: Sperm whales make
clicking sounds

rather than the long, steady
ones that we are familiar with

from other whale species.

Those clicks are easy to
translate to binary code,

so if the AI can analyse them
and find a pattern,

it may show that these whales
do have some kind of language.

Pringle: The idea is to develop
a system that is similar

to AI human language models that
produces whale vocalizations

instead of text and
then program a chatbot

that tries to converse with
whales in the wild.

Narrator: But sceptics
are quick to point out

that one of the limitations of
language models

is that they have no knowledge
about the meanings of the words

that they are using to
communicate,

they are merely
replicating patterns.

So even if the system is
successfully developed,

there's no guarantee that whale
sounds could be translated

into understandable human
language.

Aitken: And there's no way of
knowing if a whale would even

engage in conversation
with something

that isn't another whale.

Narrator: Like human beings,
animals also use a variety

of facial expressions,
body postures,

gestures and eye contact to
convey what they are feeling

or what they want.

Morgan: Cats arch their back
when they are threatened.

Dogs wag their tails
when they are happy.

Narrator: Christina Hunger
claims that Stella

is able to use her paws to
express any of her needs

with a high degree of accuracy,

often forming up to
5 different phrases

on over 40 different buttons.

Pringle: You can see in the
videos that Stella uses

her paw on different
buttons to combine words,

like "Bed, "Want",
"Come", "Eat".

Aitken: But this doesn't mean
she's actually

communicating with her owner.

She's just pressing
her paw on a button.

We don't know what
she is really thinking

because she can't talk.

Morgan: Imagine being able to
hear what an animal is thinking

or feeling just based on
its body gesture

or facial expression.

That's a dream come true for
a lot of pet owners I know.

Narrator: Not everyone is
convinced it will be reliable.

Zoologists in particular,
counter that the visual 'cues'

animals express are
far too vague

to decipher accurately into
human language.

Pringle: We may not be able
to read theirs,

but they are very adapt at
reading our non-verbal cues.

Narrator: Stella's detractors
suspect this could be the reason

why she is able to do all the
amazing things in her videos.

Her owner, Christina,
may be giving her cues,

unconsciously,
or even consciously.

Morgan: Christina is often
operating the camera,

so her facial expressions,
eye-movements

and body language
are not visible.

Aitken: So as far as we know,
she may be glancing or even

nodding in the direction
of the correct buttons.

And it's possible she's
unaware she's doing this.

Narrator: These cues might give
Christina the false illusion

Stella is
communicating with her.

It wouldn't be the
first time it has happened.

In the early 20th century,
in Berlin,

a horse named 'Hans' drew
worldwide attention

as the very first animal that
could communicate

with its owner.

Pringle: Hans was able
to answer his questions

by tapping numbers or letters
with his hoof.

But on further examination,
it was revealed

that Hans was unable to answer
any question,

if he couldn't see
his owner's face.

He was doing it by
reading the subtle signals

he was giving him.

Narrator:
It's suspected that Stella

might be doing the same thing.

Morgan: It's possible that
Christina is rewarding her dog

off-camera with treats, or that
she is only showing videos

where Stella is successful.

Narrator: Christina Hunger
denies this is the case

and claims she has never
rewarded Stella

when she was being trained.

She's so confident she taught
her dog to talk,

that Christina wrote a book
about how she did it.

Morgan: The book also
shows other owners

how they can train their
pets to do the same thing.

And some of them
swear that it works.

They feel that they can
communicate with their pets

in some basic way.

Narrator: A recent,
Amazon-sponsored report predicts

that within 10 years, a pet
translator may be ready

to be released to the public.

Pringle: This will usher in
a new age,

where pet owners can actually
talk back and forth

with their dogs and cats,
which is really exciting.

It will also allow us to
forge close emotional ties

with them and could
even save their lives.

Narrator: In the U.S. alone,

an estimated 3 million
unwanted cats and dogs

are euthanized each year--

likely because their owners
may be unaware

of the cause of their
behavioural or medical issues.

But if they could tell us what
they think,

how they feel, that need never
happen again.

Aitken: It wouldn't only
benefit pets and their owners.

AI technology can allow
farm animals

to tell ranchers
if they are sick,

and AI may also be able to
detect pain in their faces.

It could change our
entire relationship

with our four-legged friends.

Narrator: From the story
of Adam and Eve,

to modern day movies
and television,

our cultural history is filled
with fantastic tales

of animals communicating
with humans.

The question is--can
artificial intelligence

turn these fantasies
into reality?

Many believe that language is
one of the fundamental elements

of human exclusivity and trying
to communicate with animals

contradicts the
very laws of nature.

Maybe some things are better
left to the imagination.

♪♪