How AI Can Help Humans Save the Planets Biodiversity
Transcriber: Miki Murakami
Reviewer: Hani Eldalees
In 2014, I met Sudan.
The last male white northern rhino
in the world.
Just four years later, Sudan died.
Leaving two, both female,
northern white rhinos alive.
The species are effectively extinct.
And they’re not the only ones.
We’re losing biodiversity
at an unprecedented scale.
We’re in the middle of what is termed
the sixth mass extinction,
a biodiversity crisis.
And we don’t even have the scientific
and technological solution
to keep up knowing
what we’re losing and how fast.
The International Union for Conservation
of Nature Red List
is the official international organization
that keeps track of the biodiversity
of the world,
and of the 130 000 species
that they track
out of the millions that are out there.
Majority have their conservation
status as data deficient
or their population trend as unknown.
And these are iconic species
like killer whales and polar bears.
We don’t know how well they’re doing.
We can’t make policy decisions.
We can’t put the right resources
to protect them.
How many African elephants are there
and how fast are they lost to poaching?
How far do the whales go and how many
juvenile turtles survive to adulthood?
We don’t know. And these data are critical
to conservation decisions.
So how do we get those data?
There are not enough scientists to track
every animal everywhere all over the world
and not enough collars
and satellite tags to track them.
Besides to put a satellite tag
or a collar on an animal,
you have to actually capture them,
tranquilize them,
and have a vet present to monitor
the vital signs of that animal.
And even if everything goes right,
the collar can get snagged on a branch or
the satellite tag can get infected.
So this can be dangerous to the animals.
Today,
images are the most abundant
readily available source of information
about anything.
From what you had for lunch
to what animals you saw
in your backyard or in a safari tour.
Coming from scientists,
field assistants, camera traps,
as well as drones and tourists going on
safaris and whale watching tours.
There are millions of images out there.
If I could only take
those millions of images
and extract the information
about wildlife.
Well,
artificial intelligence to the rescue.
We designed algorithms
and created a platform, Wild Book,
that uses modern artificial intelligence,
machine learning and computer vision
to take these millions of images,
and find the ones that contain animals.
Find where the animals are
in those pictures,
including that baby elephant
hiding behind its mom
and figure out not only species
but down to individual animals.
Recognizing Zippy the zebra
and Joe the giraffe and Terry the turtle
and Willie the whale,
using the unique markings on
an animal body like a fingerprint.
A body print, if you wish.
The stripes, spots, wrinkles, notches,
as well as the shape of a whale’s fluke
of the dorsal fin of a dolphin.
These are unique, as every animal is.
And with information on when and
where the image was taken,
we can now use pictures.
Instead of collars and tags
to track animals.
Count them and even figure out
their social network,
who is whose animal’s friend.
This is an example page from a Wild Book
for whales and dolphins, flukebook.
And this is Pinchy, the most cited animal
in that Wild Book.
Pinchy is a celebrity. She’s a ham.
She likes getting her picture taken.
She has more than 600 sightings
around Dominica.
She lives there. She hangs out there.
And flukebook,
the Wild Book for whales and dolphins,
contains more than a million images
of almost 46 000 identified individuals.
Providing the basis
for science and conservation.
We even developed an artificial
intelligence agent,
who scours social media,
publicly-posted images and videos,
and finds the ones that contain animals.
Sends them off to this machine learning
back engine for identification
and adding to the appropriate page
the right Wild Book,
and then posting back in the comments
of the social media saying,
“Hey, at 2 minutes 46 seconds,
we found this whale shark in your video.”
“Here’s everything we know about it.”
And people respond,
“Wow, this is amazing.
How can I help?”
That…
The “how can I help?”
We engage people right were they are.
Turning their vacation videos into data
for science and conservation
with the help of artificial intelligence.
The Wild Book for whale sharks
contains data now
for more than 12 000 individual
whale sharks.
From photographs brought in
by almost 9000 citizen scientists,
200 plus conservation and
science projects,
and one very intelligent agent
from social media
that together,
that is now the foundation
for the IUCN Red List entry,
for the species.
Providing not only data
for the global population size,
but determining its conservation status
and changing it from vulnerable
to endangered.
And the population trend from
stable to decreasing.
Not because the species are doing
any worse,
but because we now know better.
We can make better decisions.
We can create better policy.
We can put the right resources
to support it.
We have Wild Books for 53 species
from marine to terrestrial,
spanning the entire globe and growing.
The technology in Wild Book was also used
for the first ever full census
of the entire species, the endangered
grevers zebra.
Using photographs from ordinary people.
Just taking pictures for two days.
For the first time, in January 2016,
hundreds of people were driving
around Kenya,
the country containing 95% of this
endangered species.
From rangers and school kids
to tourists with telephoto cameras,
they took more than 40 000 images
and the machine learning technology
of Wild Book identified all the animals,
providing the most accurate
count of the species.
So much so that Kenya Wildlife Service
said that this is how
we’re going to track
the species from now on
and do this every two years
with the event known as
the Great Grevers Rally.
So we repeated it in 2018 with more
than 1000 people, and also in 2020.
And that data became the basis
for the IUCN Red List entry
for the Grevy’s Zebra, as well as
for the conservation policy,
the Endangered Species Management
for Kenya Wildlife Service.
The associate warden of Kenya Wildlife
Service also said
this shows the power of citizen science
and machine learning for conservation.
Artificial intelligence
democratizes science,
it connects people bringing together
the pixels of individual cameras
into the global view of biodiversity.
A. I. helps create conservation policy
science and engage people at large scale
and high resolution.
And it takes
the incredible team of Wild Me,
the non-profit home of Wild Book,
as well as thousands of people
all over the world who take pictures,
annotate them and
make them ready for A. I.,
create technology and use it
for conservation.
As well as all the people
who work out there in the field
protecting the biodiversity of the planet.
And I hope you join us.
Thank you.