00:00:00
[Applause]
00:00:05
thank you Kate thank you Daria uh it's
00:00:08
such a pleasure for me to spend some
00:00:10
time with you this morning and I
00:00:11
definitely want to second what both
00:00:13
Daria and Kate said thank you for coming
00:00:16
here for those of you in the room thank
00:00:18
you for joining us online for those of
00:00:21
you who are participating this way this
00:00:24
is a really important event for us it is
00:00:27
so great to be here with all of you um
00:00:31
because right now when you wake up in
00:00:33
the morning let's be honest it's a
00:00:36
pretty tumultuous time you know I just
00:00:39
think that when you look at history
00:00:41
there are some years that witness
00:00:43
relatively little change and yet there
00:00:46
are some months where you see years
00:00:50
worth of change in what feels like a few
00:00:53
weeks and I think in many ways the first
00:00:57
three months of 2025 have brought
00:01:00
enormous change and not all of it is
00:01:03
easy for nonprofits especially when you
00:01:05
look at sources of funding and so we
00:01:08
realize what an important time this is
00:01:12
what I want to do this morning is talk
00:01:14
about where technology fits in because
00:01:17
after all as you can imagine this is
00:01:19
really a conference about what
00:01:21
technology can do for nonprofits and so
00:01:25
where I think I fit into the time that
00:01:28
we have together is to share some
00:01:32
perspective about technology and
00:01:34
especially AI what does it mean for
00:01:37
nonprofits but even more than that how
00:01:40
are we thinking about it at
00:01:42
Microsoft and let me start by putting it
00:01:46
in the context of what economists think
00:01:48
about as economists think about
00:01:51
technology there's really two types of
00:01:53
technology a generalpurpose technology
00:01:56
and a singlepurpose tool most
00:01:59
technologies in the world are in fact
00:02:01
singlepurpose tools a light bulb a smoke
00:02:04
detector a drill they do one thing
00:02:09
really really well but a generalpurpose
00:02:13
technology what economists call a G GPT
00:02:17
is a technology that basically impacts
00:02:19
the entire economy electricity is really
00:02:23
the archetype and when you think about
00:02:25
it every aspect of our economy runs on
00:02:29
electricity it changed everything and
00:02:33
interestingly what historians have
00:02:36
learned is that GPTs have really been
00:02:39
the driving force of all of the
00:02:43
industrial revolutions the first
00:02:45
industrial revolution started in England
00:02:48
in the 1700s and it was really driven by
00:02:51
iron working and the steam engine but
00:02:54
iron working more than anything else the
00:02:57
second industrial revolution really took
00:02:59
off more in the United States than
00:03:02
anywhere else and it was a combination
00:03:05
of electricity and machine tools machine
00:03:08
tools really built the modern
00:03:12
manufacturing economy and the third
00:03:15
industrial revolution is really the
00:03:17
story of our lives it started with the
00:03:21
computer chip and when it was combined
00:03:23
with software it fueled a digital era
00:03:28
now interestingly when people think
00:03:30
about these industrial revolutions and I
00:03:32
find especially when you meet with
00:03:33
people in government they always think
00:03:36
first and foremost about what it means
00:03:38
to be at the frontier of the leading
00:03:41
edge the leading sector like a GPU when
00:03:45
you're thinking about AI those things
00:03:48
matter but what history teaches us is
00:03:52
that what matters even more is not being
00:03:55
the inventor of the leading edge is what
00:03:58
economists call diffusion it's actually
00:04:03
the use of technology as quickly and as
00:04:07
broadly as possible or what economists
00:04:10
call diffusion we in the tech sector
00:04:13
call adoption and this makes great sense
00:04:16
when you think about it because after
00:04:18
all take electricity if it can change
00:04:21
every part of the
00:04:23
economy then the countries that benefit
00:04:26
most are those that use it in every part
00:04:30
of the economy and you see this in the
00:04:34
data for example this is the growth in
00:04:38
electricity consumption on a per capita
00:04:40
basis in the United States as
00:04:43
electricity consumption grew GDP per
00:04:47
capita grew as well i have slides like
00:04:50
this for basically every country on the
00:04:52
planet and they're all the same the
00:04:55
correlation is extraordinary and what
00:04:59
was true for electricity in the second
00:05:02
industrial revolution also became true
00:05:05
for
00:05:06
digitization in the third industrial
00:05:08
revolution and this is something that we
00:05:11
at Microsoft understand not only because
00:05:13
we're interested in history and we read
00:05:16
but in a sense this is actually our
00:05:20
story as a company microsoft was founded
00:05:24
on April 4th
00:05:27
1975 if you look at the calendar you
00:05:29
realize that we'll have our 50th
00:05:32
birthday a week from Friday
00:05:41
but what's interesting when I look at
00:05:42
the history of Microsoft and as somebody
00:05:44
who's been here for almost 32 years in
00:05:48
many ways we are a software company and
00:05:51
always have been a technology company
00:05:53
and have always been but we're really a
00:05:56
GPT diffusion company a generalpurpose
00:06:01
technology diffusion engine the
00:06:04
company's very first mission was defined
00:06:08
by a young Bill Gates and Paul Allen and
00:06:11
Steve Balmer it was about a computer on
00:06:14
every desk and in every home running
00:06:17
Microsoft software now what I find most
00:06:20
interesting about this first mission
00:06:22
statement is there is one word in it
00:06:26
that's used twice the word every and we
00:06:30
fast forward almost 50 years to today
00:06:34
and the font in our mission statement
00:06:37
has changed but that same word is still
00:06:42
used
00:06:43
twice now we are about empowering every
00:06:46
person and every organization on the
00:06:49
planet to achieve more
00:06:52
every is something that defines what we
00:06:55
have always tried to do as a company to
00:06:58
bring
00:06:59
technology to everyone and in many ways
00:07:02
Microsoft is one of the companies that
00:07:05
has served of as the heart of if you
00:07:07
will of this third industrial
00:07:11
revolution we'll celebrate our 50th
00:07:13
birthday next Friday it'll be fun but
00:07:16
really it's always here about the future
00:07:20
usually a year at a time maybe five
00:07:23
years at a time but I do think we're at
00:07:25
a moment when we can look forward more
00:07:27
broadly and start to imagine
00:07:30
collectively what might the second
00:07:33
quarter of the 21st century bring well
00:07:37
one thing we believe is clear even in a
00:07:40
world with so much
00:07:41
uncertainty AI really is the next
00:07:45
generalpurpose technology you think
00:07:48
about the problem on planet earth AI
00:07:51
will serve a role in helping to solve it
00:07:55
it will impact every part of the economy
00:07:58
i do think it really is the electricity
00:08:02
of our era so the real question is how
00:08:05
do we as a company really how do we as a
00:08:08
community think about what it will take
00:08:11
to ensure that this new general purpose
00:08:13
technology in fact serves the world well
00:08:17
well
00:08:18
interestingly the more we've thought
00:08:20
about it the more we've concluded
00:08:22
there's actually four critical
00:08:24
ingredients for success and I want to
00:08:26
talk about each of them briefly not
00:08:29
surprisingly it starts with the
00:08:31
technology itself
00:08:33
it turns out that every GPT every
00:08:38
general purpose technology is built with
00:08:41
a technology stack a stack of
00:08:44
technologies that need to come together
00:08:47
and you can see this from a from
00:08:49
electricity really the GPT that's most
00:08:52
familiar to most people in the world it
00:08:55
was 1878 when Thomas Edison first was
00:08:59
able to use electricity to light a light
00:09:02
bulb and then it was four years later in
00:09:06
Manhattan that for the first time there
00:09:10
was a power plant that illuminated the
00:09:13
lights in buildings the very first
00:09:15
building to light up was the New York
00:09:17
Times building
00:09:19
i think that no one uh probably imagined
00:09:23
in 1878 when Edison lit that first light
00:09:26
bulb that they were going to need to
00:09:29
build an entire tech stack but that's
00:09:32
what was required it started with the
00:09:34
fuel to power the generators in the
00:09:38
power plants and then they needed to be
00:09:41
connected with a grid that would reach
00:09:43
every building and home that was using
00:09:46
electricity there needed to be
00:09:48
transformers and circuit breakers built
00:09:50
into the grid there needed to be wiring
00:09:52
and switches and circuit breakers there
00:09:54
needed to be appliances that actually
00:09:57
made electricity useful and then there
00:10:00
were the new opportunities for
00:10:03
manufacturers for users in effect this
00:10:07
became the tech stack for electricity
00:10:09
and it created a new economy because
00:10:13
every layer of this tech stack had new
00:10:17
businesses new jobs new skills that all
00:10:20
needed to come together but what I think
00:10:23
is most interesting about this is the
00:10:27
innovation that was unleashed especially
00:10:30
at what I would call the appliance layer
00:10:33
if you go to your home when you leave
00:10:36
this or if you're watching online you
00:10:39
may be watching from home if you look
00:10:41
around your kitchen if you look around
00:10:43
your flat or apartment or house almost
00:10:46
everything you see will have been
00:10:50
invented in the first 20 years after
00:10:54
electricity took off in lower Manhattan
00:10:58
imagine what it must have felt like if
00:11:01
it was a hot summer day and for the very
00:11:04
first time you walked into a room that
00:11:07
had an electric fan or imagine what life
00:11:12
meant when you actually had a washing
00:11:15
machine or in some ways my favorite
00:11:18
imagine for the very first time walking
00:11:20
into a kitchen that had a blender i mean
00:11:24
what is this thing it's loud you know
00:11:26
and then you would realize how much
00:11:29
easier it made it to prepare dinner all
00:11:33
of these things probably felt like magic
00:11:37
and in a sense they were magical in
00:11:40
terms of the impact they had on
00:11:42
societies that could benefit from them
00:11:45
well it's no longer 1878 or 1882 now
00:11:50
it's 2025 and interestingly
00:11:53
AI is also being built on a tech stack
00:11:58
the tech stack fundamentally has three
00:12:00
layers the infrastructure layer with the
00:12:03
land and power and ships and data
00:12:05
centers the platform layer and the
00:12:08
counterpart to that appliance layer the
00:12:11
applications layer and one of the things
00:12:15
really the heart of what we're doing at
00:12:17
Microsoft from a technology perspective
00:12:20
is focusing on investing and innovating
00:12:24
in all three layers it starts with the
00:12:28
infrastructure layer which to me is just
00:12:31
extraordinary i love visiting data
00:12:33
centers at some level they all look the
00:12:35
same and at some level they're all
00:12:37
different i'm amazed by just the
00:12:40
extraordinary amount of wiring and
00:12:44
liquid cooling and the chips and
00:12:46
basically the electrical engineers and
00:12:49
electricians and the mechanical
00:12:50
engineers and the pipe fitters these
00:12:54
really are in many ways the power plants
00:12:57
of our time the critical digital
00:13:00
infrastructure on which AI relies and
00:13:04
we're building it around the world we're
00:13:05
spending $80 billion this year alone to
00:13:10
build out this infrastructure we're
00:13:12
building it in more countries than any
00:13:15
other company because you have to have
00:13:18
the infrastructure so that AI can be put
00:13:21
to work but you can't stop there it's
00:13:25
then the platform layer that makes it
00:13:27
possible to build applications that will
00:13:30
put this infrastructure to work so
00:13:32
there's foundation models like a GPT4 or
00:13:37
a
00:13:39
GPT40 coming from a company like OpenAI
00:13:42
our critical partner but we're in fact
00:13:45
seeing many of these foundation models
00:13:49
some based on large investments in
00:13:52
training some being more focused some
00:13:55
being open-source but all of those are
00:13:57
coming together they're all trained with
00:14:00
large amounts of data but what really is
00:14:03
the key at the platform layer is in
00:14:06
addition a third layer the platform
00:14:08
level services so what we're doing at
00:14:11
Microsoft is thinking and working and
00:14:14
investing and innovating in putting
00:14:16
these three pieces together there's all
00:14:19
of the platform software components that
00:14:22
we are building out because these are
00:14:25
the digital tools or tool chain as
00:14:28
software developers now refer to it that
00:14:31
make it possible for people then to
00:14:34
build applications on top that are
00:14:37
infused with the power of AI and so when
00:14:41
you really look at what we're doing and
00:14:43
I think where we so closely connect with
00:14:46
all of
00:14:47
you is the ability to support nonprofits
00:14:51
around the world and startups around the
00:14:53
world and large companies and
00:14:55
governments around the world it's really
00:14:57
all of the people who will unleash
00:15:00
innovation at the applications layer to
00:15:03
put the power of AI to work to solve the
00:15:06
world's problems
00:15:08
now all of this is like a giant flywheel
00:15:13
because in truth it all has to get
00:15:16
moving you build the infrastructure so
00:15:19
that you can train a model and deploy it
00:15:22
around the world you provide those
00:15:25
models so that applications can be built
00:15:27
on top but you need the applications to
00:15:30
take off to become popular to be used to
00:15:34
generate the revenue to keep investing
00:15:36
in the infrastructure and like a giant
00:15:40
flywheel oh there are days when it all
00:15:43
seems to work seamlessly and in symmetry
00:15:46
and most other days when there seems
00:15:49
like there's more progress at one layer
00:15:51
than another and you're constantly
00:15:53
focused if you're at a place like
00:15:55
Microsoft in identifying each area what
00:16:00
are the opportunities what are the
00:16:01
challenges what are the problems that
00:16:04
need to be solved a hundred years from
00:16:06
now people will look back and say "Oh it
00:16:08
must have been easy." Well if you're in
00:16:11
the heart of it of course you appreciate
00:16:13
that it's always hard but that's the
00:16:16
technology layer now if the only thing
00:16:19
we did as a technology company was
00:16:22
master the
00:16:23
technology we would do a quarter of what
00:16:26
is needed in order to build this new era
00:16:29
of AI so it's really combining the
00:16:34
technology with these other things
00:16:37
starting with economics
00:16:40
interestingly every tech stack actually
00:16:44
has an economic structure and that's one
00:16:47
of the really important things to always
00:16:50
think about and understand because this
00:16:52
is true of electricity it's true of AI
00:16:55
it's been true of every general purpose
00:16:57
technology if you look at electricity
00:16:59
for example what you see is that the
00:17:02
power plants are enormously expensive
00:17:05
the power grid is enormously expensive
00:17:09
but the appliances in contrast are not
00:17:13
they were cheaper to invent and
00:17:15
obviously much cheaper to manufacture or
00:17:18
buy now what is interesting is that
00:17:21
fundamental economic structure of the
00:17:24
second industrial
00:17:25
revolution is in fact being repeated
00:17:29
because AI infrastructure is very
00:17:32
expensive and this is very different
00:17:34
from say the third industrial revolution
00:17:37
when somebody like Michael Dell could
00:17:39
really make enormous progress in taking
00:17:43
the costs out of call it the hardware
00:17:45
layer and help make personal computers
00:17:48
within a few years cost a half or only a
00:17:52
third of what they cost
00:17:54
before and then the software
00:17:56
applications were built on top but this
00:17:59
era requires massive capital at the
00:18:02
bottom and the infrastructure even when
00:18:05
the opportunities at the top remain much
00:18:08
less expensive now that economic
00:18:11
structure actually translates into a
00:18:14
financial architecture because you can't
00:18:17
build this tech stack without having a
00:18:19
real vision and a strategy for the
00:18:22
financial architecture that's needed and
00:18:24
you see Microsoft not only doing this in
00:18:27
our own investments it helps explain why
00:18:30
you see headlines like the big capital
00:18:33
funds coming together for example the
00:18:35
one that Black Rockck and Microsoft and
00:18:37
MGX and then more recently Nvidia and
00:18:40
XAI are all helping to raise to generate
00:18:44
even more capital to help you know bring
00:18:48
innovation to the entire supply chain of
00:18:51
what is needed and to help invest in all
00:18:55
of this that needs to be built around
00:18:56
the world in fact interestingly enough I
00:19:00
think we're seeing this financial
00:19:02
architecture evolve before our eyes as
00:19:05
well it starts with the big investments
00:19:07
by private companies it then has this
00:19:10
private capital it has investments by
00:19:13
sovereign wealth funds and I do believe
00:19:16
that we are likely to
00:19:18
need other public funding as well to
00:19:22
fill in the gaps especially on a
00:19:25
continent like Africa where those gaps
00:19:27
are real and you reach the limits of
00:19:31
what makes sense for private capital the
00:19:34
private market to invest in and then the
00:19:38
last piece of this economic aspect is
00:19:42
really the business model it turns out
00:19:44
that you always need stable oftentimes
00:19:48
innovative business models that's what
00:19:51
sustainable success is built upon and
00:19:55
when you think about digital technology
00:19:58
it always comes down to one of three
00:20:01
business models the first is a
00:20:03
subscription it's like subscribing to a
00:20:06
magazine you pay once and you can read
00:20:09
one story a week you can read every
00:20:11
story in every issue if that's what you
00:20:13
want that is in fact M365 that we offer
00:20:18
you know somebody can buy a subscription
00:20:20
and they can use every feature very few
00:20:23
people do they can use only one
00:20:26
application but that is the way a
00:20:28
subscription works then there's
00:20:30
consumption that's the way the cloud
00:20:32
services work including Azure people pay
00:20:35
for the amount they use and in effect
00:20:38
they pay as they go and then there's
00:20:41
advertising and advertising has really
00:20:44
emerged as you all know as really being
00:20:47
at the heart of many
00:20:49
consumerbased digital services people
00:20:53
may get up and look at Instagram in the
00:20:55
morning and they'll never pay for it but
00:20:57
obviously advertising is what is paying
00:21:00
for that service to exist to me the most
00:21:03
interesting thing is just to recognize
00:21:06
business models will evolve we don't yet
00:21:09
know what they're going to look like a
00:21:11
decade from now and I think the most
00:21:13
interesting story if you look at the
00:21:15
history of generalpurpose technologies
00:21:18
is how they evolved for electricity
00:21:21
electricity grew first in the United
00:21:23
States as I mentioned and interestingly
00:21:26
enough the person who was the leader of
00:21:29
the basically the company that Thomas
00:21:31
Edison had
00:21:33
founded was visiting the UK visiting
00:21:36
England where he was born and had grown
00:21:39
up and at the time in Chicago where he
00:21:43
lived you could still walk down a street
00:21:45
this was in the 1890s electricity had
00:21:48
been around for 15 years at that point
00:21:52
and you would see some stores and some
00:21:54
homes that were clearly lit and others
00:21:57
that were still running on kerosene
00:22:00
people were still slow to adopt it but
00:22:03
this gentleman spent a weekend in
00:22:06
Brighton on the beach on the south coast
00:22:08
of England and when he arrived he walked
00:22:11
down the street and every store was lit
00:22:14
by electricity
00:22:16
and he wondered what's going on here
00:22:19
what have these people figured out that
00:22:21
we have not figured out in Chicago or
00:22:23
the United States so he found the
00:22:26
manager of the local power plant and
00:22:29
what he did was he showed him an
00:22:31
invention it was called a power meter it
00:22:35
was what was put in every store so that
00:22:38
people were not paying on a subscription
00:22:40
basis as they were in Chicago but paying
00:22:43
on a consumption basis instead it turned
00:22:47
out for the first 15 years of
00:22:49
electricity in the United States when
00:22:51
you bought a light bulb you bought a
00:22:54
subscription so that you could turn it
00:22:57
on it turned out that once the business
00:23:00
model evolved to consumption it became
00:23:03
far easier for people to go buy more
00:23:05
light bulbs and just pay the bill at the
00:23:08
end of the month i love that story
00:23:13
because I think the real lesson is not
00:23:15
just about business models it's about
00:23:19
humility every industrial revolution is
00:23:22
led by people who frankly not only are
00:23:26
really smart but they think they're
00:23:28
really
00:23:29
smart and the real lesson is that nobody
00:23:33
knows everything we all are going to
00:23:36
learn together as we go through this and
00:23:41
that is in part how we'll master the
00:23:44
economics that will be needed for
00:23:47
success now the third key ingredient you
00:23:50
might look at and sometimes people do
00:23:52
and say well I'm surprised that this
00:23:54
rate rises to the same level skilling
00:23:57
why is that as important as economics
00:24:00
and technology well it turns out that
00:24:04
skilling is the fundamental force that
00:24:07
drives the adoption and growth of each
00:24:10
of these industrial
00:24:12
revolutions it makes sense because if a
00:24:15
new technology is going to be used
00:24:17
across the economy the skills to put it
00:24:20
to work need to be mastered across the
00:24:24
economy so why did ironwork take off in
00:24:28
England it wasn't just because George
00:24:30
Watt had invented the steam engine there
00:24:33
it was because England at the time had a
00:24:35
system of technical institutes and
00:24:38
apprenticeships that taught people in
00:24:40
the evening how to master iron working
00:24:43
and so iron working spread more quickly
00:24:46
the US benefited from this amazing
00:24:48
coincidence when it came to
00:24:51
electricity and machine tooling because
00:24:55
in
00:24:56
1862 during the Civil War in the United
00:24:59
States Abraham Lincoln had championed
00:25:02
and then signed into law what became
00:25:04
known as land grant colleges the federal
00:25:07
government granted federal land to the
00:25:09
states to create land grant colleges i'm
00:25:12
sure some of you have degrees from them
00:25:15
and it was all started to really advance
00:25:18
an understanding of agricultural science
00:25:20
and agricultural engineering but it led
00:25:23
to this new discipline in the United
00:25:26
States called mechanical engineering and
00:25:29
so because the United States had
00:25:31
mechanical
00:25:33
engineers they were able to figure out
00:25:35
how to put machine tools to work to
00:25:39
change manufacturing across the economy
00:25:41
and then when the industry standardized
00:25:43
they were able to go even faster and the
00:25:46
same thing was true in the third
00:25:48
industrial revolution
00:25:50
interestingly employers invested in
00:25:52
training of employees in the 1980s and
00:25:56
computer science departments absolutely
00:25:58
swept the the nation in the United
00:26:01
States all the major colleges and
00:26:04
universities created these computer
00:26:06
science departments so the US had more
00:26:08
people who could code on a per capita
00:26:11
basis than any other country and people
00:26:14
have learned over time that this need
00:26:16
for skilling is not just deep it is
00:26:20
broad one of the best illustrations of
00:26:23
this was what the electricity industry
00:26:26
realized in the
00:26:27
1950s they had built out power plants
00:26:31
but electricity was not being used as
00:26:33
widely as the industry hoped so the
00:26:36
industry got together and said you know
00:26:38
what we need to do we need to help the
00:26:41
American public actually just learn what
00:26:44
a new generation of appliances can do in
00:26:47
their fir in their homes we need to
00:26:49
bring this into people's homes using the
00:26:52
power of television we have to find
00:26:54
somebody who can connect with the public
00:26:57
and help them learn about this in an
00:27:01
interesting and even enjoyable way
00:27:04
so they found this fellow who was
00:27:06
working in Las Vegas at the time he was
00:27:09
an actor he wasn't wellknown at the time
00:27:13
but he had this natural ability not only
00:27:16
to communicate but to
00:27:18
connect his name was Ronald Reagan
00:27:23
and so Ronald Reagan with his wife Nancy
00:27:26
started to come to everyone's home every
00:27:30
Sunday evening when GE would host
00:27:34
basically a theater a play that evening
00:27:38
but before it began Ronald and Nancy
00:27:41
would show off the latest appliance in
00:27:45
their home they ended up with so many
00:27:47
appliances in their home that they
00:27:50
needed to put in place an additional
00:27:52
electrical generator just to power all
00:27:54
of it but it worked and it probably was
00:27:59
indispensable in the rest of his career
00:28:01
including becoming president of the
00:28:03
United States it shows how skilling
00:28:07
needs to connect with people i think one
00:28:10
of our great opportunities and
00:28:13
challenges as a company as a community
00:28:15
as an industry really as a planet is to
00:28:19
think about AI skilling at scale and
00:28:21
make it one of the great opportunities
00:28:23
and causes together for the next decade
00:28:27
and two to come and it really starts by
00:28:31
understanding thinking about what are
00:28:33
the skills that people need to learn and
00:28:35
we should recognize these are early days
00:28:39
we don't yet know exactly what we're
00:28:41
going to need but we do think there are
00:28:43
at least three categories there's
00:28:45
fluency just learning how to use AI
00:28:47
learning how to use a co-pilot or chat
00:28:50
GBT or in other everyday software
00:28:53
applications that is what we are doing
00:28:56
that's what we are doing in partnership
00:28:58
with many of you it is a little bit like
00:29:01
what Ronald Reagan did we just haven't
00:29:03
hired Ronald Reagan yet the next is AI
00:29:07
engineering i think this is the future
00:29:09
of computer science the real question is
00:29:13
what will a computer science degree look
00:29:15
like a decade from now will it be an AI
00:29:18
science degree what will people need to
00:29:21
learn in order to really create AI
00:29:25
applications there's a good chance it
00:29:27
will involve less code because AI is
00:29:31
getting very adept at coding but an
00:29:33
enormous amount of design and probably I
00:29:36
think more multiple disciplines because
00:29:39
people need to really master all of the
00:29:43
impacts of AI and how to make them more
00:29:45
useful and then there's what we think
00:29:47
about as AI systems design if you're an
00:29:50
organizational leader how do you
00:29:52
understand your workflows what we would
00:29:54
think of as your business processes
00:29:57
which ones are most likely to benefit
00:29:59
from the application of AI how do you
00:30:02
measure the success how do you take
00:30:05
people through cultural change and in
00:30:08
effect every country is going to need to
00:30:11
develop its own national AI talent
00:30:14
strategy assessing their economy looking
00:30:16
at the different sectors where are their
00:30:19
real needs for skilling and what type
00:30:22
where is there the ability to partner
00:30:25
often with nonprofits and definitely
00:30:28
through the education sector to build AI
00:30:31
fluency for everyone and build out all
00:30:34
of the other skills that are needed all
00:30:36
of these things will need to come
00:30:38
together to master skilling and then you
00:30:41
have the final piece it's what
00:30:44
historians and political and social
00:30:47
scientists call social acceptance
00:30:51
it turns out that broad technology
00:30:54
adoption requires social acceptance it
00:30:58
sort of makes sense when you think about
00:31:00
it but in the 1980s social scientists
00:31:03
went to work sociologists in particular
00:31:06
and they asked a very important question
00:31:09
why did some technologies take off and
00:31:11
get used broadly when others did not and
00:31:14
it turned out that their studies showed
00:31:17
it always came down to two factors first
00:31:20
something needed to be useful unless it
00:31:22
was useful people wouldn't use it that
00:31:25
seems obvious but the other is that it
00:31:28
had to be
00:31:29
trusted and I think it's this element of
00:31:32
trust that is also at the heart of what
00:31:35
we need to advance when it comes to AI
00:31:39
and that's what we're doing and as we
00:31:41
think about trust we see these four
00:31:44
elements it's there's security there's
00:31:46
privacy there's digital safety i'd say
00:31:49
especially the protection of children
00:31:50
and others and there's this discipline
00:31:53
called responsible AI that has emerged
00:31:56
in the industry and has really spread
00:31:59
around the world including through often
00:32:01
times new laws and
00:32:03
regulations and just as there's a tech
00:32:06
stack for the technology itself there's
00:32:09
sort of a stack that we're building for
00:32:12
AI governance there's an architecture i
00:32:15
think it starts with the internal
00:32:17
policies at tech companies we have now
00:32:20
corporate standards in these places for
00:32:23
these topics we train engineers we have
00:32:25
engineering tools we have compliance
00:32:28
systems and the same is true for
00:32:31
customers whether they're nonprofits or
00:32:34
companies companies or governments but I
00:32:37
think on top of that we're seeing emerge
00:32:39
industry standards the standards are
00:32:42
critical because they define best
00:32:44
practices and as best practices emerge
00:32:47
there's the foundation for say the
00:32:49
domestic policy ultimately the
00:32:52
international policies that are needed
00:32:55
this too needs to come together in order
00:32:58
for AI to spread around the world and
00:33:01
because this is 2025 and not
00:33:05
1882 we have to add an element to this
00:33:09
aspect for social acceptance called
00:33:11
environmental
00:33:13
sustainability because it does turn out
00:33:16
as everyone knows that those big data
00:33:19
centers that you saw the pictures of run
00:33:21
on a large amount of electricity which
00:33:24
is why we're so focused and why we
00:33:26
remain so committed to achieving in 2030
00:33:31
the goals we set for ourselves in 2020
00:33:34
to be carbon negative by the end of this
00:33:36
decade and what that means is reducing
00:33:39
our emissions from electricity and from
00:33:42
the use of greener steel and greener
00:33:44
concrete and greener fuels and the like
00:33:48
and then engaging in carbon removal it
00:33:51
is what has made Microsoft the largest
00:33:55
corporate investor on the planet of the
00:33:58
removal of carbon from the environment
00:34:01
so that's how we achieve we believe
00:34:03
social acceptance what I think is most
00:34:06
interesting about this is to be truly
00:34:11
successful you actually have to do not
00:34:13
one of those four things you have to do
00:34:16
all four of them at the same time and I
00:34:19
do think that what perhaps
00:34:21
differentiates us from say other tech
00:34:24
companies more than anything else it's
00:34:27
the fact that we are working so hard to
00:34:29
master all four of these things together
00:34:32
and you're going to see new initiatives
00:34:35
new innovations and new investments in
00:34:38
the coming months in the next year in
00:34:41
all four of these areas
00:34:43
i would then
00:34:45
conclude by just offering a thought
00:34:48
about what the history of technology
00:34:51
teaches us i'd first start with the
00:34:54
cautionary tale it goes back to
00:34:57
electricity i actually believe that the
00:35:01
diffusion of
00:35:03
electricity was both remarkable and at
00:35:06
the same time represents the single
00:35:08
greatest technology tragedy in history
00:35:12
what is the tragedy it's the fact that
00:35:15
today we come
00:35:16
together 143 years after that power
00:35:20
plant started operating in Manhattan and
00:35:23
there are still 700 million people on
00:35:26
planet Earth that
00:35:28
tonight won't be able to use electricity
00:35:31
to light a light bulb because they don't
00:35:34
have access to it it's 43% of the people
00:35:38
who live in Africa i think that the
00:35:41
number one cause of the great global
00:35:45
divide between the global north and the
00:35:47
global south the economic divide that
00:35:49
has afflicted so many now generations of
00:35:53
people's lives has fundamentally been
00:35:56
based on whether they lived in a place
00:35:58
that had access to
00:36:01
electricity we need to do better
00:36:08
let's face it if it takes 150 years to
00:36:12
bring AI to the world then we will have
00:36:15
failed our goal I think needs to bring
00:36:18
AI to the world not in 150 years but in
00:36:22
15 that is the opportunity that we have
00:36:25
to think a new and act differently and
00:36:28
then what it asks us to do is think
00:36:32
about how we work together
00:36:35
because you're at
00:36:37
Microsoft I would start with the purpose
00:36:39
of a company sati Nadella our CEO likes
00:36:42
to quote a professor named Colin Mayer
00:36:45
who said "The purpose of a company is to
00:36:47
find profitable solutions to the world's
00:36:51
problems." I fundamentally believe that
00:36:54
every healthy society is built on a
00:36:57
healthy three-legged stool there are
00:37:01
governments that obtain money by taxing
00:37:04
it and they spend money to solve the
00:37:08
world's problems they create the
00:37:11
fundamental social infrastructure needed
00:37:14
for a civilized society and then you
00:37:17
have nonprofits that raise money to
00:37:21
solve the world's problems and then you
00:37:24
have companies that sell products and
00:37:28
services to earn a profit to solve the
00:37:32
world's problems and what is most
00:37:35
interesting in the world today in my
00:37:37
view is that while people often spend
00:37:40
their time talking about how we're
00:37:42
different what we don't talk about is
00:37:45
how we do our best work not just as
00:37:48
separate organizations but in
00:37:50
communities and countries when we all
00:37:53
compleiment each other because that is
00:37:57
what we
00:38:01
do so in some
00:38:04
we know our role well it is to pursue
00:38:09
profitable solutions to the world's
00:38:11
problems mostly by building out that
00:38:14
infrastructure layer building out the
00:38:16
platform layer so you can go to work as
00:38:20
can we with applications that will make
00:38:24
the world a better place in essence our
00:38:28
role is to serve you at Microsoft we
00:38:31
serve the world's
00:38:33
nonprofits so you can serve the world
00:38:37
and when we do that together even in a
00:38:41
hard day a hard month or a hard year we
00:38:45
make the world a better place thank you
00:38:49
very much
00:38:52
[Applause]
