Gutenberg's invention of the printing press led to a new economy, which created a new sector of businesses, industries, and jobs. Generative AI is providing a similar opportunity today. In this episode, Brad Smith draws on the lessons from the printing press and supporting industries to illustrate how different technologies are coming together to create a new AI economy. He explains the critical partnerships between countries, government regulators, and tech companies - building the necessary infrastructure required to run generative AI to developing the skills people need to use it. Brad also discusses how Microsoft’s AI Access Principles foster both innovation and a path for new businesses by ensuring access, fairness, and responsibility.
Gutenberg's invention of the printing press led to a new economy, which created a new sector of businesses, industries, and jobs. Generative AI is providing a similar opportunity today.
In this episode, Brad Smith draws on the lessons from the printing press and supporting industries to illustrate how different technologies are coming together to create a new AI economy. He explains the critical partnerships between countries, government regulators, and tech companies - building the necessary infrastructure required to run generative AI to developing the skills people need to use it. Brad also discusses how Microsoft’s AI Access Principles foster both innovation and a path for new businesses by ensuring access, fairness, and responsibility.
Click here for the full transcript.
Brad Smith: I'm Brad Smith and this is Tools and Weapons.
On this podcast, I'm sharing conversations with leaders who are at the intersection of the promise and the peril of the digital age. We'll explore technology's role in the world as we look for new solutions for society's biggest challenges.
Brad Smith: Hello everybody. In this episode, I’m sharing the keynote I recently gave in Barcelona at the Mobile World Congress. I focused on what we call “the new AI economy” – in effect, a new sector of the economy that’s coming together with a number of different technology pieces, that all need to work together to enable AI to flourish. One of the interesting things I find as I meet with people around the world is, understandably, we’re still at a very early stage of people even understanding what all these pieces are. What is an AI data center? What is a foundation model? What does it mean to try to bring them to the world?
I addressed another critical question as well. How should a big tech company like Microsoft build this infrastructure, make it available to the world, and do so in a way that can give people confidence? We announced I shared new AI Access principles, principles that will follow at Microsoft. Fundamentally, it's about three things, access, fairness and responsibility.
Fundamentally, it’s about three things: access, fairness, and responsibility. The new AI economy, Well, listen here – up next on Tools and Weapons.
Brad Smith: Good afternoon. It's my pleasure to be here to represent Microsoft at this year's Mobile World Congress. I think this brings an important group of people together at a critical moment in time. I'm going to talk today about the new AI economy, the new sector of the economy that AI is building.
And I come here on the 12th day of almost two weeks spent in Europe, where we've had the opportunity to announce more than $5 billion of investment, money we'll spend just this year and next in Germany and Spain to construct new AI data center infrastructure. It's not just about bringing billions of dollars of investment to countries in Europe. We've announced a new training program to bring AI skills to more than a million people, new services based on AI to empower network operators, a new partnership announced today with Mistral AI in France, and new initiatives to focus on the use of AI to advance cybersecurity and AI safety itself.
Brad Smith: What does this all amount to? That's what I'd like to talk about today.
Interestingly enough, despite the rapid pace at which invention and events are flowing, we're still in an extraordinarily early stage of really thinking about what generative AI is creating. And I find as we meet with people, in some ways, all of us in our industry have an important need and opportunity to help the world learn about what we together are creating, about this new sector of the economy. And whenever you have something very new, it helps by comparing it to something that people are familiar with. In many ways, when I think about AI, I think it's most interesting to compare it to the printing press, literally an invention of almost 600 years ago.
Brad Smith: Why? Well, in some ways, I think AI is the most important invention for the life of the mind since the invention of the printing press. Think about what the printing press enabled people to do. To write, to formulate their thoughts, to put them down, to share them so that other people could read, and by reading, it enabled them to learn.
In so many ways, AI is a tool to help people think, to create, to reason, to share, to learn, much like the printing press. But the comparison does not end there, because the printing press was not just about inventing something new.
Brad Smith: There's a second aspect that was important as well. It was about the diffusion of technology, meaning the need to get it into people's hands, to put it to work. And it's fascinating to look at what happened in the 50 years in Europe, in the second-half of the 1400s. The printing press was built with movable type by one person, Johannes Gutenberg, in Mainz, what is now Germany. And over the course of five decades, it's spread from city to city, from country to country.
Brad Smith: But what is also interesting is it wasn't just an invention that was created. It was an entirely new economy. If the people who were involved in the technology of the printing press had had something like Mobile World Congress, they would have come together, perhaps in Mainz, perhaps in Barcelona, and they would have done then what we do today. They would have looked at this economy and they would have said there's a tech stack just like technology people do today. And when they looked at the tech stack, they would see that it started with the printing press, but fundamentally it required so many more parts of the economy to be built and come together.
There needed to be people who produce paper and ink. There were people who were book binders, book publishers, booksellers. They all made it possible for writers and readers to flourish, and flourish they did. It took this entire economy to make the modern world.
Brad Smith: In fact, what is most interesting is when they all came together and succeeded together, books exploded. If you look at the world in 1460, in the prior 1000 years, only 1000 literary works were produced – sorry, 8000 literary works in 1000 years. But soon 8000 literary works were being produced just every five years. It wasn't just that new works were being produced. The printing press made it possible for them to be provided in great number.
By the end of the century, 20 million copies of books had been created and distributed and even read. What that made possible is almost everything we know about European history that followed. It led to the Renaissance and then the Reformation, to the Scientific Revolution, to the Enlightenment, to the spread of democracy itself. That is what a great technology can do.
Brad Smith: When people come together and build a successful sector of the new economy and use that technology to do good for others, I think it should inspire us.
So as we come together here in 2024, we too are creating a new sector of the economy. We're creating the new AI economy. And one of the first things we need to do to think about ourselves to share with others is the tech stack that we are building. It is a tech stack that starts with two things the world often takes for granted, electricity and connectivity. But unless those two come together, nothing else is possible. But then we need the AI chips. We need the AI infrastructure. You see, all of the parts of the AI economy that need to come together.
Brad Smith: The creation of foundation models, the tooling that makes it possible for people to use those models and create applications, the app stores that are needed to distribute AI powered applications to consumers, and ultimately the people we all serve. Like readers and writers, there are now users and developers that are bringing this technology to people around the world. The truth is we all need to succeed together, and certainly at Microsoft we see our role in our particular niche and niches as one that is inherently and indispensably connected with everyone else. More than anything else, we're spending more money literally than anyone else on the creation of the next generation of digital infrastructure, of AI data center infrastructure that literally is now being built around the world. It's an amazing thing to visit a data center today because it's extraordinary. It has now oftentimes dozens of buildings.
Brad Smith: We don't build data centers. We build data center campuses, and those buildings have millions of servers inside them. Fundamentally, this is where we need to spend money to make it possible for companies to train AI models, to deploy those models and run AI enabled applications to serve people's needs around the world.
But increasingly we find people ask, how are we going to run this infrastructure?
Brad Smith: In many ways, we're doing what no one ever imagined doing in something like the age of electricity. Thomas Edison invented the light bulb, but no one ever thought that they could take on the role of diffusing electricity quickly enough by building power plants around the world. As a result, electricity diffused slowly. So slowly, in fact, that today, 142 years after the first building in New York City was illuminated by electrical power, there are still 700 million people in the world that woke up this morning with no access to electricity at all.
Brad Smith: Our goal is to build a future that puts technology in people's hands quickly and equitably so the entire world can prosper and benefit from what AI has to offer. But to do that, we're going to need to be principled as a company, and that's what I want to share this afternoon. Today, we're announcing and launching what we're calling Microsoft's AI Access Principles. These are the principles that we've been assembling as the technology has been moving forward. These are the principles that define how we at Microsoft will govern the technology we are creating.
It starts with a set of goals, 5 tenets or goals that really are guiding how we think. The 1st is the role and even the responsibility that we have to enable AI innovation and foster competition so that it serves everyone and not just a small or single company. It means that our obligations start with the law itself. We live in a world where new technology and especially AI laws are being created quickly and we need to respect and abide by them. But that is the floor.
Brad Smith: That is where our obligations begin. We have much that we need to build on top of that. Fundamentally for us, it's all about partnerships. We're not vertically integrated. We don't have an app store for consumers. We rely on others for chips. We partner with others to create the world's greatest foundation AI models. If ever there were a time where there was a need for partnership to advance a new technology, it is here and it is now. But unlike many prior eras, especially of digital technology, I think partnership in this new AI era, partnership for this new AI economy, requires that one think broadly.
We need to partner not just with other companies like those in the room but with entire communities and countries as well. We need to partner with each of our customers who are creating AI enabled applications. And we need to do so by being proactive, by being responsible, by not waiting for problems to grow, but instead moving quickly to solve them. These are the goals that guide us.
Brad Smith: We've thought about these goals, and we're sharing today 11 principles. These principles really fall into three categories. They're about access, they're about fairness, and they're about our broader societal responsibilities. But first and foremost, perhaps they are about access, about providing broad access to other companies, communities and customers. Broad access, so first, others can train their models and deploy them on the AI data center infrastructure we're creating. This is illustrated by the announcement we made just two hours ago.
Brad Smith: What I think really stands for a new day and a new era for Microsoft support for the development of technology in Europe. It's the multi year partnership we just announced with Mistral AI, the leading AI company in France. And under this partnership Mistral AI will now be able to train and deploy its leading models in our data center. But it's not just training and deploying models. It's about helping people bring their models to customers, to software developers, so they can use them to customers, so they can buy them.
And through our Models as a Service initiative at Microsoft, Mistral AI, effective today, has its models available to reach the world through all of the data centers that we're operating. We're focused not just on proprietary software, not just on our partnership with Open AI, as critical as it is, but with a wide number of companies. And in fact, as of today, we have almost 1600 models running in our data centers, 1500 of which are open source models. It shows how we as a company need to focus on proprietary and open source models, companies large and small. That is what it means to provide access to the world for the data centers that we're building to construct.
Brad Smith: But it's not access alone that matters. It's also incredibly important, we appreciate, to treat people fairly, because when you're running the critical data centers on which people rely, they need to have confidence in how we will interact with them. We've given a lot of thought to what we think matters to different groups of people. We think that a commitment to fairness starts by publishing the API, the application programming interface, so anyone who wants to create a model or create an application that runs on a model has access to the public versions of the APIs, the same APIs that we use ourselves. That's a third principle.
It's an important step in ensuring that everyone is treated fairly. But we also recognize the critical role that network operators play, the investments that network operators are making in building out the infrastructure for the future and we recognize, through the GSM Association's Open Gateway initiative, the big opportunity ahead.The opportunity to take the data that network operators have, as you know, data about where people are, data about who people are, because they're authenticated through their accounts.
Brad Smith: Data that operators have about the quality of service that an individual has at a particular moment in time. And to take this data and give network operators the ability to connect directly with the people who are creating AI enabled applications. They can build better applications for people. They can build applications that protect people, including from the risk of fraud. And they can do it in a way that will open up a new source of revenue that will help network operators grow, and in so doing, help them modernize the network and the connectivity on which we all rely.
Brad Smith: But it's not even those two things alone. We then need to think about the commitments that we need to make as a company to ensure that the people creating AI enabled applications can reach consumers broadly. That's why we have decided not to follow in the footsteps, in this case, of great companies like Apple and Google, but companies that have built app stores that have become gateways through which all commerce and apps must flow. We've created an app store, we call it the Azure Marketplace, but we give applications developers a choice. They can use it or not. They can decide how they want to get their technology into the hands of customers. And as we give them that kind of choice, we sign up, I think critically, for the need to show developers the respect they deserve.
If they are training a model on our infrastructure, if they are deploying it on our infrastructure, we recognize that their data is their data. We will not access it and use it to compete with the companies that are relying on our infrastructure.
Brad Smith: And finally, we are committed to giving people the choice even to leave if they want to take their model or if they want to take their data and move it on to another cloud. It's our responsibility to enable them easily to do so. When you put those five commitments together, we think this is an important step to making it clear how people could rely on Microsoft and how we want to support them. Then there's the final thing that we're committed to doing. It is to addressing our broad societal responsibilities.
More than ever, people and governments around the world have a high expectation for us, indeed a high expectation for all of us. And that means that we need to continue to take important strides forward. First, to protect the physical and cybersecurity of these data centers around the world and all the data and technology inside, even against a more hostile set of ransomware, criminal enterprises, and even nation state adversaries.
Brad Smith: It means we need to continue to invest in a high standard for responsible AI, safety architecture that protects people's safety and security, the privacy of their information. And we are doing this in ways that not only safeguard and put guardrails around our own services like our Copilots, but we're doing it in ways where we make this safety architecture available as a service for open source developers so they can use it if they wish, rather than have to spend their time and money creating it themselves.
We recognize that ultimately the success of this new economy requires not only companies like all of ours who are represented in this room. More than we've seen in a long time, we need to get out and skill people around the world so that they have the capability to use the power of AI to advance their careers.
Brad Smith: And finally, we need to do all of this in a way that is sensitive to the environment, that conserves the use of water, that recycles it, or oftentimes runs without any water at all. We need to do it in a way that brings online renewable and carbon free energy, more energy even than we may consume in our data centers. We need to do it in ways that put the power of AI to work to protect the environment.
Ultimately, when you put these 11 principles together, we think that they're an important step forward in making clear what we feel we need to do to earn and sustain the trust of people, of customers, of companies, of countries around the world. And we recognize that the first step is almost never the last step. But the best way to get going is to get started, and that's what we're doing. As much as anything else, perhaps more than anything else, we think this is a time for us to look at each other and appreciate the role that we each are playing. We all need to work together. That is what will define our success.
And we need to work together because at the end of the day, we all need to succeed the only way we can, by succeeding together.
Thank you very much.
Brad Smith: You've been listening to Tools and Weapons with me, Brad Smith. If you enjoyed today's show, please follow us wherever you like to listen.
Our executive producers are Carol Ann Browne and Aaron Thiese. This episode of Tools and Weapons was produced by Corina Hernandez and Jordan Rothlein. This podcast is edited and mixed by Jennie Cataldo with production support by Sam Kirkpatrick at Run Studios. Original music by Angular Wave Research.
Tools and Weapons is the production of Microsoft made in partnership with Listen.