IR Fireside Chat Series: Advancing AI Sustainably and Responsibly

Brett Iversen: Welcome everyone. I'm Brett Iversen, vice president of investor relations at Microsoft. Thanks for joining us for our latest installment in our series of videos, focusing on strategic areas that are top of mind for our investors. Today's discussion will focus on advancing AI sustainably, and responsibly, issues that are top of mind for all investors, including those focused on our environmental, social, and governance initiatives, also known as ESG.

We brought together three of our key leaders to answer your most frequently asked questions. We have Melanie Nakagawa, CVP, and chief sustainability officer, Noelle Walsh, CVP of cloud operations and innovation, and Natasha Crampton, Vice President and chief responsible AI officer. As always, please reach out to our investor relations team directly with any feedback you might have after you view the video. With that, let's kick things off. So first of all, thank all of you for being here. I'm excited for the topic. Just to ground the investor audience to start with, maybe each of you could share a little bit about your area purview, core commitments that you're focused on, just to give people a grounding. Melanie, maybe you could kick that off.

Melanie Nakagawa: Sure. So, at Microsoft sustainability started most recently in 2020 when we announced our bold and ambitious goals to be carbon negative, water positive, and a zero waste company, all by 2030, and then by 2050, remove all of the emissions since our founding from electricity in 1975, our historic emissions. But one of the things four years into this journey is we know that we have to extend our sustainability impact beyond our four walls. It has to extend to our customers, and our partners by providing them with technology and innovation that helps them on their sustainability journey as well.

But then to step back even further, it's also really important as the chief sustainability officer that we also deliver an outsized impact, a global impact. How we work with other companies, and partners, and other organizations to really invest in the type of infrastructure, and solutions we're going to need, not only for us, but for so many others around the world. So, that's things such as advancing sustainability policy, using our Climate Innovation Fund to invest in innovations into new technology solutions, as well as working with other partners around their sustainability commitments as well. We know it's going to take all of us, so the purview is quite large, but it really extends, and starts at home.

Brett Iversen: Perfect. Natasha?

Natasha Crampton: Thanks, Brad. It's great to be here with you all. Look, my job in a nutshell has two parts. The first part is really internal facing part, where it's all about operationalizing our six AI principles. We adopted those back in 2018, and they are that AI systems should be fair, they should be reliable and safe, private and secure, inclusive, accountable, and transparent. Now, all of our AI systems, including all of the Copilots that we've made available in the last year, have these principles built into them by design, because we've built processes and systems to help our engineering teams carry out that work on a day-to-day basis.

The second part of my job is a more external facing one. It's all about advancing the new laws, and norms, and standards that we need on AI. As much as Microsoft recognizes its own deep responsibility in the space, and tech-led initiatives can help pave the way here, the reality is we can't do it alone, and governments around the world are really advancing the state of the art when it comes to new laws, and norms and standards. So, we participate in those conversations around the world.

Now, across both of those priorities, I'm focused on sharing our learnings, empowering external parties with the knowledge that they need to have to develop shared guardrails, and of course learning from others that are invested in this work, as well. Ultimately, our goal is to not only build AI systems responsibly, but really help our customers and partners in their journeys, as well, and give policy makers the knowledge, and experience that they need in order to develop a shared set of guardrails so we can realize the benefit of this technology.

Brett Iversen: Perfect. Thank you. Noelle?

Noelle Walsh: Hello, Brett. Yes. Now, I'm responsible for designing, building and operating our data centers, the infrastructure that powers the Microsoft Cloud. Currently, we operate at over 60 Azure regions, across 34 countries globally, more than any other data center provider. I also oversee our substantial renewable energy team, and carbon capture team, ensuring our expansion aligns with our sustainability commitments for the company.

And whether or not we realize it, we all rely on cloud services, be it for online banking, travel arrangements, schooling, medical, connecting with friends, and online shopping, and much more. But beyond our day-to-day personal lives, global enterprises and businesses depend on cloud, and more than ever before. Over 95% of the Fortune 500 are hosted by our Microsoft Cloud. So, cloud consumption was already driving exponential data center growth, and AI has been a growth multiplier on top of that. So, our Microsoft infrastructure is playing a key role in meeting the AI moment, and without that infrastructure, we could not host AI.

Brett Iversen: Well, you touched, all of you, a little bit on this in your first answer, but maybe a little bit more on the importance of each of the areas that you all lead, and how you operationalize such broad areas across a company as large and diverse as Microsoft. So I'd love to hear from all you on this. Maybe Noelle, you can start on this one.

Noelle Walsh: Yeah, just to touch on that briefly, Brett, I mean, Microsoft data centers are vital for our cloud services. Operating them at high availability each and every day, and every day of the year is critical to our customers' success. So that is a very strong focus area, as is keeping pace with the exponential growth that they demand going forward. And I believe not just what we do is important, but how we do it, that we do it with high personal safety on the job, sustainability on our design and operations, physical security across our data center fleet, and being good neighbors in the communities in which we operate.

Brett Iversen: Melanie?

Melanie Nakagawa: So, for Microsoft, our sustainability also as I mentioned, needs to extend beyond our four walls. It can't just be for us alone, but it has to be for all the customers, and partners we get to work with every day. So, obviously our journey starts with meeting our own commitments to be - access to be carbon negative, water positive, a zero waste business, and to extend that to so many others, as well. But an example on the carbon negativity side, the way we operationalize that goal is to work closely obviously with Noelle's team, who's focusing really hard on how do we ensure that we're transitioning to more and more carbon-free sources of power, and how do we support the growth of these carbon-free sources of power in the key markets in which we operate?

But it's also about how do you invest in durable carbon removal solutions as well? Because while we are focused on reduction, part of our commitment to negativity also involves the removal of carbon as well, from high durability projects. So, this is areas that is still a nascent market, but we're really working hard to build that market, and where it can exist.

And then this last year we also recognize it's a shared responsibility. These sustainability commitments are not just sustainability commitments for the CSO, but it's one that's shared across the business. And we were fortunate and privileged that over the last year, senior leaders, including Noelle and her team, have leaned in from across the business to develop over 80 interventions that are going to help us toward that carbon negativity roadmap, and how we execute those. There's also ways that we've gotten to work with other senior leaders, including the board, and presenting this plan for how we can lean in, to meet some of these roadmaps and be able to bring that to the environmental and social public policy committee of the board.

Brett Iversen: Helpful. Thank you. Natasha?

Natasha Crampton: Well, if I can just zoom out for moment, and speak to why we do this responsible AI work at Microsoft. Look, I believe strongly that if AI is advanced responsibly, it can create positive ripples around the globe. But like other technologies that come before it, that is not a foregone conclusion. AI does have the potential to either narrow, or widen existing economic and social divides. And so, the reason we have a deliberate strategy to govern this technology is so that we can realize the potential benefits of this technology in a way that benefits everyone. And of course, that is very much in line with Microsoft's mission.

The other thing that we are guided by is that we know from past waves of digital transformation that people don't use technology that they don't trust. And so, it's core to our practice of responsible AI that we are trying to build justified trust and confidence in these exciting new technologies. So, in terms of how we've operationalized this over the years, we've been working on responsible AI for eight years now, and that has been building on Microsoft's legacy as a company, which is now at almost 50 years of bringing new technologies into the world responsibly.

People are at the very core of our responsible AI program. Across the company is the responsibility of all employees to practice responsible AI. In fact, at this point, we have more than 400 people working on responsible AI across the company, with more than half of them working on it full time. And when we look across our full employee base, 99% of our employees completed the responsible AI module when it came to our standards of business conduct training last year. We've adopted a responsible AI standard that sets out the practices and the processes that teams need to follow in order to uphold those six principles that I mentioned earlier, and we've also built tools to help make it easy, not just for our teams within the company, but also our customers, and partners to operationalize responsible AI as well.

Similar to sustainability, responsible AI is not a compliance activity at Microsoft. It's a deeply embedded part of our culture. It's a part of our thoughtful, and deliberate approach to making AI systems available to our customers, and partners. Like Melanie, we regularly report on our progress to the environmental, social and public policy committee of the board, and we have built leadership forums within the company to make sure that we have the highest levels of accountability. Fundamentally, this is all made possible by the ecosystem of people that we have working on responsible AI within the company, and also outside the company. It builds on this strong foundation of research, the expertise of our policy experts, and engineering teams who are building scaled processes and tools to make this a reality on a day-to-day basis.

Brett Iversen: All of you sound very busy, so I appreciate the time today. So, Noelle, this next one's for you, because one of the biggest questions we get from investors is, how are you as a company scaling our infra in line with the great customer demand that we're hearing for our AI solutions? And so, how is that going? What's our confidence in continuing to do that? So what would you share on that space? Because I'm sure you spend most all of your waking hours thinking about this. What would you share with our audience?

Noelle Walsh: Oh, I could share so much, but yes.

Brett Iversen: Yeah. Yeah. That's right.

Noelle Walsh: A significant portion of my time is dedicated to scaling our operations to meet the AI growth. And I would say within Microsoft, and my organization, we're very excited to be part of that at this moment in time. We announced new investments across 12 countries within an eight-month time period, to expand our cloud services, and bring AI to so many countries. And most nations are recognizing the strategic importance of having their own data center infrastructure, and that's becoming increasingly apparent as we continue to grow.

So, I'm dedicated to ensure not only that we build to this demand going forward, but that we are ready to operate it, and with that high degree of availability, because this is an unprecedented infrastructure growth expansion. So, it involves comprehensive recruiting plans, to have the right people in the right places globally. So we partner with local communities, colleges, governments, and businesses, to hire the right people to operate our data centers.

In addition, working across Microsoft, we're focusing on upskilling our communities, and vast states, and countries, to ensure everyone have the right skills, and as you mentioned as well, to be part of this new AI economy. So, we have a strong focus not then just on hiring, but onboarding and training, emphasizing innovation, and bringing technology into the world. So, it is a challenging and very dynamic time, and with the right leadership teams, and strong industry partnerships across our ecosystem, we are poised to meet the moment successfully.

Brett Iversen: Thank you. And maybe that's a good way to build on that, Melanie, I could go to you. And so we're doing all of that at pace, at scale, but still as a company needing to meet the pretty ambitious ESG commitments we set for ourselves. So, what would you share with the audience on how that's going, and how we think about those two things together?

Melanie Nakagawa: No, it's a great segue between the work that we get to do together all the time. And so, one of the things to just ground ourselves is that, so while we all know that AI computing is more resource intensive, the advances in AI technologies not only have the potential to accelerate sustainability progress, but are actually critical to achieving not only Microsoft's carbon negative goals by 2030, but frankly, the world's goals to be a net-zero economy by 2050. So, an example of this to share for all of you is, AI has been able to advance significant opportunities to accelerate this progress. One example has been how did we use AI to accelerate clean energy transmission? Noelle was talking about the build, and to be ready to operate. One of the key areas that we've been seeing progress with AI is how AI is actually helping existing transmission lines bring on new renewable energy capacity at a speed and scale that's much faster, and that's so critical to this.

It's also helping discover new types of batteries, and chemistries that help get new types of long-duration battery storage. That's also really important for intermittent sources like wind and solar. And we have to be really thoughtful. Like with any new tool, we also must be really thoughtful with how we deploy that. And so, we're really thinking about the resources that the new builds are going to require from energy, and water, and carbon. And the good news has been we've been taking action across each of these areas. One of the key ways we think about is how are we designing for sustainability from the design, the construction, to the operations of our data centers? On the design side, areas like water and carbon efficiency, or energy efficiency, and the design of our data centers. When it comes to construction, how do we start to accelerate markets for new low-carbon building materials, such as steel and concrete that we put into these?

We made an investment last year in a company called H2 Green Steel with our Climate Innovation Fund, and they're looking to have the first large-scale commercial green steel facility that can reduce steel emissions by over 90%. If we can do that, and build like that, we're actually taking control over the types of embodied carbon that's in our data center footprint. And then on the operations side, once you build these data centers, how are we ensuring that we operate them sustainably? Through Microsoft's, and Noelle's team in particular, they've actually achieved over 30 gigawatts of clean energy power purchase agreements to power many of these assets. And that's a great way of operationalizing our sustainability commitments when it comes to our data center growth. So, as Noelle said, we are poised well for actually being able to do both, be able to grow our AI economy, while also doing that sustainably.

Brett Iversen: So, there's a word you said, thoughtful, that is a good word. Natasha, to go to you, so we are scaling at kind of unseen pace, which is exciting. We're doing that, keeping ESG commitments in mind, and delivering against those. But we also want to be thoughtful that we're thinking about AI responsibly as well, and that we're doing our part to make sure it's used as intended. What can you share in terms of how your team's thinking about that? Because that's a question we get, as well.

Natasha Crampton: Absolutely. And just in May this year, we were excited to release our very first Responsible AI Transparency report, and that was really reflecting our first year of shipping AI systems at scale, generative AI systems at scale, and it encapsulates many of the learnings that we had in meeting that demand. One of the core things that we've been doing is working to align our approach internally with a government-developed standard, the NIST AI risk management framework, which has been developed here by NIST in the US, and well-received around the world.

Now, at its core, this is a risk management-based approach. So, across four key functions, govern, map, measure, and manage, we are monitoring the life cycle of the AI systems that we're building. So, just to give you a bit of a flavor of what that looks like in practice, governance of course is a lot about making sure that we've got the right people in the right role. So, being clear about roles and responsibilities, bringing diverse perspectives to bear on the AI systems that we're building. That training that I mentioned earlier, the accountability through leadership fora that I also mentioned earlier.

Map is about understanding the new zones of risk for generative AI systems. So, we use techniques like impact assessments, like security threat modeling. We've adapted the security practice of rear teaming to expand to responsible AI risks more broadly, as well. That's all part of our mapping set of techniques that we use. Measurement is a key way in which we understand the behavior of AI systems, and make sure that they're aligned with our principles. So, we've developed metrics to assess the performance of our AI systems, and to check that those mitigations that we are putting in place, actually doing the job that we need them to.

Managing is about making sure those mitigations are in place, and making sure that they're effective before, and during, and after launch. So, we make sure, for example, that we do have incident response processes and the like, to make sure that if there is something that we didn't catch ahead of time, we can quickly respond to it, and address it in a more systemic way.

To give you a flavor of what this all looks like in practice, one of the case studies that we highlight in our Transparency Report is about Copilot Studio. This is a fantastic platform, whereby people who do not have deep AI experience, or years of coding experience can actually build their own copilots. One of the risks that the team identified early on is that it's very important in certain factual contexts to make sure that the Copilot is responding with data that is well-grounded in the sources that it's based on.

So, they identified this risk, they then developed metrics to measure for it, and then we used some mitigations that we've built, and scaled out to our customers through the Azure OpenAI service, to make that the Copilot was actually returning in-context results that made sense based on those data sources. There are many more examples in the Transparency Report, and I would encourage our viewers to dig in there. I think the key takeaway is that we have built this durable risk management approach that scales across the company, and that evolves as well, as we identify new risks, and adopt new practices to address them.

Brett Iversen: Well, I'd love to stay with you for one more, because another question that we get a decent amount from investors is with AI, and the number of solutions, and the newness of the innovation, how does Microsoft think about regulation? Are we for more global regulation on this topic? Are we worried about that? What are we doing? What are we hoping for? I know my point of view on that, I'd love to hear the official point of view, and how you think about all that.

Natasha Crampton: Sure. So, while as a company, we have a deep sense of responsibility for the systems that we make available, as I said earlier, governments need to be involved here. We cannot do it alone. And so, we are supportive of regulation in this space, and we've been actively contributing to the conversations about how that regulation should be formed. And so, one of the things that's very heartening to me about the policy conversations that we've been having is that they've started so much earlier for AI than what they have for many other digital technologies. And for me, this gives me a sense of optimism that we're on a path to putting in place the right set of shared guardrails for this technology via regulation.

So, in addition to the industry led efforts, we've already seen domestic and regional pieces of legislation start to land, such as the European Union's AI Act. At the international or global governance level we are also starting to see some of the pieces of the jigsaw puzzle come together there as well. Just as a couple of recent examples, there was an AI Seoul Summit in Korea in May, and there we had 10 countries around the world, plus the European Union, agree to establish a network of AI safety institutes, and they're really going to help with advancing that measurement science that we believe is critical to safely deploying AI systems in the world. And in fact, we now have two resolutions at the United Nations level, all 193 member states coming together to commit to a forward path for the safe development and deployment of AI, and also on capability building, so that the benefits of AI can be shared broadly.

Recently as Microsoft, we made a contribution to this conversation about global governance by making available a book, which has lessons, and insights for the global governance of AI drawn from other domains. So we had the privilege of working with about six experts across a range of different global governance disciplines. We had climate experts, experts in the IPCC, we had nuclear energy governance experts, we had civil aviation experts, financial services, and a comparative governance expert, who came in and shared lessons about the history, and the evolution of those governance mechanisms, so that we could have a discussion about what they mean for AI.

Across these case studies. We saw a few things. First, the importance of striving to reach consensus built on a strong scientific foundation, like what we have been building in the climate space through the IPCC. We also learned about the importance of facilitating cooperation across countries that don't necessarily see eye to eye all the time. We need to leave room for the evolution of governance as the technology changes, and we also need to take into account the different needs of countries around the world.

Now, for Microsoft's part, we believe if we step back, and we think about three big objectives for global governance, we need to manage globally significant safety and security risks. We need to guide towards regulatory interoperability, so different systems across the world can interoperate with one another. And third, we need to ensure inclusive progress, so that we can all enjoy the benefits of AI. So, if we can achieve those three goals at the global governance level, then we believe that we'll be on a good path to achieving the mission that Microsoft has for AI more broadly.

Brett Iversen: That's great. I mean, getting global regulatory alignment is really hard. So, it's excellent to hear that we've been able to make this much progress this quickly. So, thank you for that. Melanie, so if I reflect a year ago, when we were chatting, I still remember some of the education you gave me on scope one, two and three emissions. And if I'm remembering right, I remember scope three being, at least in my mind, the hardest, because I think of it as kind of the indirect bucket, with supply chain upstream, all the way through to how customers use our products, and making sure we're doing our part to lower, and control that bucket of emissions as well, which seems very hard to me. I'm sure you would agree. What's the latest on how we're thinking about that, what we're focused on? What would you share with our audience on that one?

Melanie Nakagawa: Well, just like last year, you are wowing me with your knowledge of scope three, once again. So, first to just ground ourselves before... Because not everyone has your knowledge of all of Microsoft's emissions, although this panel does, I'm sure, as well. So first, every year we put out a sustainability report that lays out in great detail really where we stand as a company when it comes to our emissions, and the progress we got to report this past year in our 2024 sustainability report was the incredible progress we've actually made on our scope one and two emissions. That's the direct emissions. A lot of that work that Noelle talked about around data centers, that's what gets represented in those emissions.

And this past year that scope one and two declined by 6.3%, which is just a great kudos to the type of work, and the progress, and scaling that Noelle has been talking about. This includes things like the announcement with Brookfield Asset Management, and Brookfield Renewables for over 10 and a half gigawatts in a clean energy power purchase agreement that actually identifies more clean energy to power these data centers, to really reduce that footprint and carbon from our data centers. But you're absolutely right. The area that we tend to focus a lot around is in scope three, and that's that value chain, the supply chain. It makes up over 96% of Microsoft's reported emissions. That's why we're so focused on it. We need those numbers to go down, and that requires partnering with our suppliers, and our supply chain to make that happen.

In this past year's report, that scope three emissions was actually 30.9% higher than our 2020 baseline. So, that's been an area of focus that since we met last year, what are we doing to actually reduce emissions in that category? That category includes things like the embodied carbon in our data centers, so that's the servers, the hardware components, the racks, the cement, the steel, the fuel, all that is the key drivers of our scope three footprint. And so we started talking about the kinds of actions we're taking. We're investing in green steel businesses, we're investing in low carbon concrete, we're partnering with semiconductors.

But to simplify, I'd say there are three areas over the past year that we've been really targeting for our scope three reductions. The first is to increase our measurement around where it comes from, and the efficiency in our operations, like our data work. The second area is partnering with our suppliers, and our supply chain to help them on their sustainability journey, getting them access to cleaner types of steel, or cement, helping them get access to clean energy, and really engaging with them on what are the commitments we need to make, and how they ladder into our goals.

And the third key area is public policy. This is so vital, because in order for Microsoft to meet our commitments, our manufacturers, our suppliers, and countries located all around the world actually need to be on cleaner energy grids, and be accelerating their progress to sustainability commitments. As we saw last year in the climate conference in Dubai, we recognize the world's not on track to meet the Paris Commitment, Global Commitments on Sustainability. So, it's even more critical that companies like Microsoft and our that we really remain steadfast in our commitment to drive and accelerate the sustainability progress in our supply chain, and value chain, to help those around the world accelerate their progress as well.

Brett Iversen: Thank you. Noelle, when we get questions from investors, a lot of the excitement is on the innovation around the AI solutions we're bringing to customers, and partners directly. There's also, I would presume, the ability to use some of that same innovation enabled by AI to help us think of different ways to scale, and bring these services to market. I know we're early in the journey so far, but what would you share in terms of how you and your team are thinking about that? Anything you would highlight thus far, on the innovation side? On the delivery side?

Noelle Walsh: Yes. No, absolutely. It's not a question of growing, or sustainability, or innovation. It's and, and, and. So we're focused on growth, and doing it in the right way. As mentioned, it's personal safety, it's sustainability, physical security, and community partnership. So, I'll start by saying we're on track to hit our company 2025 goals with respect to renewable energy. As mentioned, we've acquired over 30 gigawatt globally that will be operationalized by the end of next year to meet the equivalent of 100% global. We are also working towards our higher order goal in 2030 of achieving 100% consumption, 100% of the time, matched by zero carbon purchases. So, when the sun doesn't shine, and when the wind does not blow.

And we have a lot of the solutions in place. We're working with a number of industry leaders, and innovators on leading edge technologies, such as supporting helium with fusion technology, which would be wonderful for society. So, we are designing, building, and operating to world-class data center standards, continually adapting day-to-day operations to minimize our impact on the environment.

We design for high power efficiencies, which for example, means that we predominantly use air to cool the servers across our data centers, and more than half of the year, and those higher temperature humidity times of the rest of the year, we'll use adiabatic cooling with very limited water. So, we're highly cognizant of power efficiency, and water efficiency. In countries, just to give some example where it is feasible, we use our air off gas, our hot air, to heat local district heating systems. That's our plan on one of our campuses, in Denmark, and two in Finland. And in Sweden, we've deployed first of a kind battery systems replacing generators, and that's just to mention some examples. With the new AI chips, be it for AI supercomputers or inferencing, we're designing for much, much less water, with a liquid cooling system.

In addition, we have power efficiency teams working across data center infrastructure, my colleagues in server hardware, in engineering services, to ensure that we are operating at high, overall high power efficiency, if we look at our entire system each and every day. On the embodied carbon materials, we're adapting to timber in a number of cases. We use recycled materials in our concrete, and we are working with suppliers who use green steel. So, we are making strong year-over-year progress, despite our massive growth, and the challenges ahead. I will say in addition, when it comes to silicon, and our server hardware, so we have first of a kind, we call them circular centers, where we can deconstruct decommissioned servers, and recycle over 90% of those servers for reuse. So, in close, it's not an either or, it's an and, and, and, and I think what an exciting moment in time.

Brett Iversen: That's a great list. I like the, "When the sun doesn't shine and the wind doesn't blow." That reminds me of budgeting when I hear that.

Noelle Walsh: Okay.

Brett Iversen: It resonates. I won't give you much of a break. We've been kind of talking to all of you about some of the harder things in each of your areas, which is important. I'd love to close, and hear from each of you on this one. What are you most excited about? And some of this you might've touched on already, but I'd love to hear, I think the audience would appreciate that, as well, and Noelle, maybe you can start us on that.

Noelle Walsh: Oh, sure. No, with I think our new large language models, AI has swiftly evolved to become a mainstream technology, outpacing human capabilities in tasks ranging from image recognition, to language translations, and so much more. I mean, ChatGPT reached 100 million users within three months alone, which was phenomenal. And the use of generative AI has nearly doubled in the last six months. So, I believe that AI holds the promise of revolutionizing society by tackling global challenges. It is a special moment in time to unleash human creativity, and the possibilities are endless, and I personally feel very privileged to be a part of this journey.

Brett Iversen: Yeah, you've a front-row seat, given your job. Melanie, how about you?

Melanie Nakagawa: I think this optimism is exactly the type of feeling that I also have when it comes to where we're moving in this new AI era. And I guess, much like my colleagues, we're in this unique position to get to see some of the most incredible innovations that AI can offer. And so, last year we put out an accelerating AI - Accelerating sustainability with AI playbook that really has - It's just chock-full of examples and case studies of how we're already using AI today, to accelerate sustainability progress. So I do encourage those to take a look at the playbook, because it really lists out everything from, I mentioned, how do you get more of this clean energy onto transmission grids, but it also includes use cases around biodiversity, and ecosystems, as well as areas like water, and how are we helping to address actually more water replenishment?

Water is a finite resource. That's why Noelle shared all the incredible work that we're doing to be more water efficient, and be mindful of our water consumption. But at the same time, we also need to be replenishing more water where we can. And so, there's incredible use cases where we can use AI sensors to detect leaks faster, and help provide more replenishment to key basins in which we operate. And so, some of these use cases around - In terms of sparking the imagination, and how we're using to actually accelerate sustainability progress today is one incredible area.

And then the second incredible area is around how we're using it around from the human component. So, for sustainability leaders, these are, we're looking at complex data sets, we're looking at complex information, and we just scale our teams, and our capacity, and AI has this incredible superpower of being able to optimize, and measure, and predict, and forecast really complex systems. And so, it's actually helping scale the sustainability workforce, so that more and more employees, not only at Microsoft, but elsewhere, can make sustainability part of their jobs, and really understand how to leverage all this data that we have to deliver more sustainability insights, to create better decision-making, and get everyone else, including ourselves, on our sustainability journey, and make that progress faster.

Brett Iversen: Absolutely. Natasha?

Natasha Crampton: Well, in addition to all these amazing use cases that Noelle, and Melanie have referred to, I think I draw a sense of my optimism from how we are approaching this set of new technologies. I'm encouraged that when I speak to customers about the responsible AI program that we've built at Microsoft, they are thirsty to do the same thing in their own organizations. They want the tools, the playbooks, the practices. They're very concerned to do the right thing. And I think particularly since the ChatGPT moment, we've seen this broad societal conversation unfold about what we want AI to do and achieve in the world, and also what we don't want AI to exacerbate, or amplify. And I think this broad and inclusive societal conversation, including with policymakers, is what we need in this space.

My overall feeling is that there's a real recognition, and a commitment in practice to a shared responsibility for AI. Melanie mentioned this earlier, but it is going to take not just the developers of the technology, the deployers of the technology, the adopters, the whole communities around it, to actually lean in, and make sure that we are steering this technology towards the potential that I very deeply believe it has. So, I think with that shared commitment, and with this commitment to making sure that we are building AI by design, in a way that is sustainable, that is responsible, I'm confident that we can secure all these amazing benefits, and the potential that's outlined by all of the use cases that Noelle and Melanie outlined.

Brett Iversen: It's amazing how much progress there's been already in all of your spaces in this short amount of time, at least seemingly to me. So, thank all of you for the time today. I really appreciate the time, and the insights. And thanks to everyone for watching. Advancing AI sustainably and responsibly is one of the top priorities across the company, and we're glad we had some time to talk with you about it today. So, thank you.