Neutralizing the Impact of Artificial Intelligence on Climate Change
[Applause]
i made sustainability my default choice
in the year 2018
and my vision for the year 2030 is to
neutralize the impact of artificial
intelligence on climate change
seven years one or three days 15 hours
12 minutes
and 6 seconds the clock here is from
manhattan
which shows you the amount of time left
to reverse climate change
seven years one not three days 15 hours
12 minutes and 6 seconds let that sink
in
on 25th of september 2019 approximately
6 million people
across the world protested against the
lack of action
and ownership by global leaders on
climate change
yet there are still a lot of people in
the world who do not believe climate
change is happening
the survey here is done by the yale
school of environment on united states
of america
which shows you that 12 percent of the
people do not believe in global warming
while 47 believe that it’s not going to
affect them
but climate change is real whether you
believe in the science of it or not
and it will affect you even though you
do not live in the cities that are
currently facing wildfires hurricanes
or are going to get submerged by the
year 2050 in water due to climate change
the world health organization estimates
that if climate change remains unchecked
then from the year 2030 to the year 2050
you will have about 2 lakh 50 000 deaths
per year
in fact there’s also research which
shows that climate change affects your
mental health
so how do we stop this a lot of people
believe a way to decarbonize
is to decomputerize technology does have
a massive impact on climate change
technology such as artificial
intelligence intel internet of things
has its own carbon footprint a team at
stanford came up with a tool called
green ai
which shows you the amount of carbon
emissions generated by an artificial
intelligent algorithm
according to the tool a natural language
processing algorithm which is widely
used
in alexa city and other customer service
chatbots
generates about 1 400 pounds of carbon
emissions
which is equivalent to a person taking a
round trip from new york to san
francisco
so does this mean we stop making
innovations and technology
technology has been the one thing that
has kept us together in this kobe 19
pandemic
technology is the way i’m communicating
my thoughts with you
technology is the method that we are
using to online conference our office
colleagues and even video chatting our
friends and family
or even to get your most basic amenities
from online grocery shopping
in fact technology is the very same tool
that we use to even spread awareness on
climate change
each tweet you post generates about 0.2
grams of carbon emissions
and there are about 500 million tweets
which go out in a day
short well let me tell you how each
tweet gets stored and processed in data
centers
these data centers are used for almost
everything on internet
from google searches to instagram posts
to even netflix and amazon
to process the tweet and to share it
with the rest of the world
these data centers run 24 7. now this
energy to run these data centers comes
from burning of fossil fuels
to give you an idea of the intensity of
this in the year 2016 it was reported
that the data centers across the world
consumed about 416.2 terawatts of energy
which was more than the energy
consumption of britain that year
hence making data centers responsible
for 2 percent of global emissions
research shows that if data centers
continue the way they do
then by the year 2040 they will be
responsible for 14
of global emissions which is equivalent
to the emissions of united states of
america right now
so how do we fix this is there a way for
us to neutralize the
negative impact of technology by using
technology for good
the answer is absolutely yes one of the
most common solutions for this problem
is to go green
that is to use renewable energy sources
to power your data centers
instead of fossil fuels but renewable
energy sources comes with its own
challenges
challenge number one they’re
intermittent in nature that is
they’re not constant 24 7. which leads
us to our challenge number two
can these intermittent sources be able
to supply
energy for 24 7 running data centers
also just shifting towards renewable
energy is not enough
the energy pyramid here shows you
exactly that
there is no point of just shifting
towards renewable energy we need to
ensure that the systems
are also energy efficient and have
energy conservation in picture
let me share with you three solutions
where artificial intelligence algorithms
can be used
solution number one artificial
intelligence algorithms such as load
monitoring can be used with your sensors
to manage and track your demand of the
data centers
this ensures that the systems in the
data centers work more efficiently
hence ensuring less amount of energy
consumption
in fact google has already implemented
this and seen a 40 percentage drop in
their energy consumption
solution number two since we knew
renewable energy is not constant in
nature
we need to ensure that the demand of the
data centers is always met by the supply
this can be done by predicting your
supply and demand by using time series
forecasting
solution number three artificial
intelligence algorithms can be used to
schedule data processing across
data centers in the world for example
consider these two data centers
data center one is in usa while data
center 2 is in asia
during the morning in asia where solar
energy is at its peak
the data centers here can run on solar
energy and process the data
while during the night this processing
can be shifted to usa
where it’s still day and solar and
energy there can be used
each of these solutions are already
implemented but
in small scale each of these solutions
can reduce the carbon impact by minimum
20
to 60 percent not only are these
solutions
environment friendly but also budget
friendly since renewable energy is much
cheaper than fossil fuels
so now that we know these solutions
exist what do we need to make this
happen
we need two things one for the present
and one for the future
let’s start with the present first like
bill gates said
we need climate denominators we need
researchers
entrepreneurs scientists engineers
who understand the correlation between
climate change and technology
and create solutions like this now for
the future
we need reducing carbon impact as one of
the core values
while developing anything new in
technology this not only applies for
hardware but software as well
every algorithm that we generate we need
to ensure that the processing takes the
least amount of energy
hence ensuring low carbon emissions with
this i leave you with my vision not just
for 2030 but also beyond it
progress in technology and mitigating
climate change needs to go hand in hand
neither one of them should happen at the
cost of the other
you