


11 Oct 2026 · 47m · 23 chapters

Microsoft president Brad Smith argues the UK should lead in setting and regulating AI safety rules, warning that AI is advancing faster than regulators can respond. He proposes “safety/emergency brakes” (kill switches) for autonomous/critical systems, mandatory incident reporting when AI agents “break loose,” and government sign-off/independent evaluation for frontier models before release. He also discusses data centers: public concerns, energy costs, and whether data centers should be built in the UK versus abroad.
Brad Smith is President of Microsoft (33 years in tech; previously senior at Microsoft). He references AI Safety Institute (UK) and regulators like the UK Competition and Markets Authority (CMA). He cites figures such as Sam Altman and Dario Amodei (Anthropic) and UN Secretary-General António Guterres.
self-regulation is necessary but insufficient; safety rules should be as routine as for planes/milk/washing machines; neighbors shouldn’t bear electricity-price burdens; communities should get legally binding commitments.
emergency brakes in trains/lifts/school buses; aircraft certification; milk/carton inspections; Microsoft’s dispute with the Obama administration over privacy; the CMA/Activision Blizzard controversy; Vice President J.D. Vance’s H-1B/green-card criticism (Smith disputes the numbers).
Written by AI. May contain mistakes.
Brad Smith shares insights on AI's impact on jobs and regulation challenges.
“Because today we're speaking to the president of Microsoft, Brad Smith.”
Brad discusses the importance of AI safety regulations and lessons from social media.
“Can I start by quoting you in a blog on Microsoft's website from a few years back?”
Three key safety measures for AI systems are proposed by Brad Smith.
“You pick the number of bullet points, but what should be the areas of focus?”
Discussion on the need for international standards in AI safety and regulation.
“And let's link three and one together, because I feel like they are kind of connected on it.”
Analyzing the balance between self-regulation in tech and the need for oversight.
“When does safety become an issue that is so important for a technology that you require the creator of the technology to satisfy a government regulator that it is safe before the product is released to the public?”
Learn about the necessity of a regulatory framework for AI to ensure safety and accountability.
“It sounds like from what you're saying to me without putting words in your mouth, you don't think self-regulation actually goes far enough.”
Understand how public anxiety shapes opinions on AI and data centers in communities.
“But it's clearly an issue around the world as well about the public thinking, hang on a minute, this is going too fast for me actually as well.”
Explore the discussion around who should pay for the energy costs associated with data centers.
“We're putting in place people and budgets and signing up where communities want to legally binding commitments.”
Delve into how the political environment affects the AI industry and its regulations.
“And then that gets reallocated so that they pay for a portion as well.”
Learn about the implications of global data policies on national security and data governance.
“Well, it's been a long time since I've heard somebody say that the people of the United Kingdom should be subject to the laws of Ireland.”
Discussing the increasing political implications of AI technology and its impact on relations with governments.
“So let's have the conversation and let's understand that there may be room for differing views and different choices.”
Addressing accusations regarding Microsoft's hiring practices and visa usage amidst layoffs.
“And then there are days that we disagree.”
Explaining the truth behind employment changes and the hiring of foreign workers.
“With the same spirit, look, I think Vice President Vance is encouraging us and others to hire more Americans.”
Discussing the balance between hiring local talent and incorporating international expertise.
“And the other thing that I would say is when you look at the data, including of all of the companies that apply for visas, we are among the very highest in paying the highest wages.”
Exploring the UK's potential as a leader in technology post-Brexit and its competitive advantages.
“Is that an opportunity for us for you to hire more people here?”
Discussing the need for a balance in creating opportunities for locals while welcoming international talent.
“And I'll start with, hey, my wife and I lived here for the last four years before I joined Microsoft from 1989 to 93.”
Evaluating the UK’s environment for investment and growth amid regulatory changes.
“That has sort of defined the state of many countries and even our world over the last 30 years.”
Exploring how AI regulations can shape the economy and ensure safety and fairness.
“There's opportunities for more speed and permitting, especially around electricity and the like.”
Learn about the need for AI regulation to protect jobs and the economy.
“They don't see that leak over to model companies so that the model companies use that to train models to, frankly, outperform their customers.”
Discover the optimistic perspective on how AI can enhance job satisfaction.
“that you insisted on having a personal computer in your first job back in the 80s.”
Hear a compelling call to action to shape the future of work with AI.
“Now, then let's talk about your second question.”
Discover the optimistic perspective on how AI can enhance job satisfaction.
“That's what I pursued when I entered the workforce in the 80s and said, give me a computer on my desk.”
Understand the rising importance of soft skills in the age of technology.
“And frankly, look, look at the folks that you sometimes see in big jobs, the big boss in some of the tech companies in our industry.”

Get the gist, explore an idea or ask about something said.
On Big Boss Interview, what are the key points of this episode?
Automatic transcript. May contain errors.
0:00Will Bain: This BBC podcast is supported by ads outside the UK.
0:30Will Bain: Syrup. Spice up fall at Whole Foods Market. Balance your trading strategy by adding futures. CME Group helps you manage risk and capture opportunities in all market environments. Capitalize on around-the-clock access to highly liquid global futures and options markets across all major asset classes. Visit your online broker and get started. See what adding futures can do for you at cmegroup.com forward slash podcast. Derivatives are not suitable for all investors and involve the risk of losing more than the amount originally deposited and any profit you might have made. This is not a recommendation or offer to buy, sell or attain any specific investment or service.
1:18Will Bain: Hello and welcome to the Big Boss interview podcast from the BBC where we talk to those business leaders shaping the way we live and work. And we haven't done many bigger, I guess, than our guests today in the past 60 or so episodes of Big Boss Interview Podcast. Because today we're speaking to the president of Microsoft, Brad Smith. And as a result, as you can imagine, lots in there about AI, the way it's shaping how we work, what it's doing to jobs. But, of course, critically, given everything that's been talked about through the summer, including by former staffers at companies like OpenAI and Anthropic Safety.
1:54Will Bain: safety and brad smith really interesting on that about how he thinks the uk could lead the world in terms of setting and regulating some of those safety standards but also about his own fears and concerns about the pace at which regulators are rushing uh in his mind not quickly enough to catch up with what's going on in arn we talk data centers as well obviously a big political football particularly back home for brad smith in the united states and well you might remember this in a conversation in 2023 with Sean around regulation here in the UK. Unfortunately, I think it's bad for Britain. The business community, the investment community, and the technology sector around the world have been following this case.
2:35And the strong message the CMA has sent is not just to surprise everyone who fully expected this acquisition to be approved, but to send a message that I think will discourage innovation and investment in the United Kingdom.
2:50Will Bain: If you don't remember that, of course, that was that enormous row when Microsoft were trying to buy the video games maker Activision Blizzard and he thought it was really hampering UK investment, how kind of involved regulators here were getting. So it would be interesting to see where he thinks we are now given there's been a lot of work around regulation to try and boost growth across the UK. So absolutely loads for us to get to. So this is it. my conversation with Brad Smith, the president of Microsoft. Brad Smith, thanks so much for being with us on the Big Boss Interview podcast. Thank you for having me.
3:21Will Bain: Can I start by quoting you in a blog on Microsoft's website from a few years back? You're talking about AI safety here, an issue I know that you think is important. It's not enough to focus only on the many opportunities to use AI to improve people's lives. This is perhaps one of the most important lessons from the role of social media. We've had a bit of a pickle trying to put regulations back on social media after the event, haven't we? So what lessons are you hoping we've learned, that we are learning perhaps at the moment? I think there's perhaps two lessons that are most important. The first is to recognize that any technology is both a tool and a weapon.
3:59And we need to look at the downside as well as the upside and manage it. The second is don't wait until it's too late. Don't wait until it's impacted in an entire generation of people and recognize that you need to put a lot of different things together to manage that downside well, including laws and regulation. And I think that's exactly where we are right now when we think about the safety of AI.
4:25Will Bain: I was going to ask you exactly that. You don't think we're past some of that point because I think some people are worried that we are, that it's hurtling too quickly and already it might be difficult to catch up with. I'm thinking of people, Helen Toner, who we've had on the program before, runs an AI safety institute at Georgetown now, was on the board at OpenAI, a company you invest in. People like that who are sounding the siren here a little bit. I think there's a lot of times in life when you look at a problem and you wish you could have started earlier. I often joke inside Microsoft. I come in and I look around and say, did any of our engineers build a time machine last night?
5:01No. Okay. Well, then let's get started today. And I think that's the way we should look at AI safety. Let's get started today and really start to do more.
5:09Will Bain: Shape it then. Keep it simple for people. But what? You pick the number of bullet points, but what should be the areas of focus? Well, I would say three things to focus on. First, put a safety break or emergency break on any AI system that is, say, controlling and on autonomous device, critical infrastructure. This is this so-called kill switch idea. Yeah, I would call it a safety or emergency brake. Because it's not got the word kill in it. I just think it captures that you may want to slow things down or turn them off, but you want to have that choice. And let's just remember we have circuit breakers in our homes.
5:45There are emergency brakes on our trains and buses. This is a common feature. Even every lift that we get in has that kind of safety mechanism. So let's build it into AI and let's put it not only in the companies that are building and running the models, but in the cloud infrastructure where a lot of times these are being run or in other places so that we can address that. That's number one. Number two, let's have what's called incident reporting. We're all used to this now. For example, if there's a privacy breach, if somebody's been hacked and your data is unfortunately let loose, you're notified so that you know what's going on.
6:26In the same way, if a company builds AI agents and they break loose, as we've seen recently, if they break in, there should be a legal obligation to tell the victim what has happened. I think that...
6:40Will Bain: A legal obligation. I think there should be a legal obligation. I think those are two things that are relatively straightforward. I think the third issue is really the biggest, the broadest, and maybe the most important. When should we have not only outside evaluators testing, say a frontier model, the most powerful new technology? When should those evaluators work for a government agency? When should they have to sign off on something before the product is released? And right now around the world, especially say in the United States, that idea makes people nervous. And there's good reasons to be very careful and thoughtful.
7:21But just remember, right now, your washing machine is subject to more safety rules than AI. Every carton of milk that you see in the grocery store is subject to safety rules and inspection. Or think about the airplanes we get on. You can get on a plane in London. You can fly to Athens or New York, either way, that plane has been certified so that it is considered safe by the aircraft regulators in both countries. We know how to do this for other products. We should build on that knowledge and start to think about how to apply it to AI.
8:01Will Bain: Well, let's do that. And let's link three and one together, because I feel like they are kind of connected on it. Because the big question I have when we're running out number one is who is doing that? Interestingly, it's really, at its best, a bit of a combination. You often get your best work done, especially on something like a safety break, for example, when the industry itself starts to develop common standards. That means identify the best practices, build out the standards, and then the regulators then take that standard and they make it a legal obligation. Take something like a school bus.
8:41I don't know how it works in the UK. I'll tell you in the United States, there is an industry standard for an emergency break for school buses. And then every municipality actually defines the law for its municipality, but they all basically use the same standard. So you start to put that kind of architecture together. I think one of the things, especially in the realm of evaluation, safety inspection that one has to remember is somebody always has to go first. Somebody has to be at the forefront.
9:13Will Bain: And it's got to be international, right? And in an early stage technology like this where people are racing, right, that people are still trying to establish who is where in this, that presumably is much more difficult than say with something, I don't know, where we're coming to a consensus after the event. Very much so. And if you go back to something like commercial aircraft, not surprisingly, that started in many ways with the United States. It started even before the Second World War ended with AI. One of the most interesting things is I actually think the country and the government that is in the best position to advance AI safety standards and even rules is here.
9:54It's the United Kingdom. It's the AI Security Institute. It started with the name AI Safety Institute. You need capability to do this well. You need to have trained people. And the UK was at the forefront. It started to build this up. The US started and then it slowed down. And so the UK has kept pace much more with the technology. In the world of technology, in the tech sector, most people that I know would say, if you ask them, name the country that is the leader in having this technical capability, they would say, oh, it's very simple. It's the United Kingdom. They are the best. Now that the UK is the best, I think the question is how to build on that.
10:40Will Bain: And how to compel people to do that too, right? It was Anthropic, one of your rivals who refused to very recently allow one of their models to be tested by the UK AI Safety Institute. If it's opt in, opt out like that, it doesn't matter who's leading the way, doesn't it? It's pointless. Ultimately, you do get to this interesting question. When does safety become an issue that is so important for a technology that you require the creator of the technology to satisfy a government regulator that it is safe before the product is released to the public? I think that's the conversation we need to start having.
11:20I don't think one should say, oh, this is Monday now, Tuesday, let's make it the rule. But let's start to have that conversation. The other thing that I think is so interesting about the United Kingdom is it's in such a good position. It is highly regarded around the world. It is a bridge across the Atlantic and across the English Channel. it's a voice of calm and reason and that doesn't mean that it can lead with bluff and buster which is bluster which is not a uk tendency anyway um build on that um and there are other countries including 29 countries that recently signed a letter yeah but i can think of one country
12:05Will Bain: in particular that might not be very keen on that and it's the one that you've joined us from today and certainly your president i think would have a pretty big problem wouldn't he if the uk was in charge of regulating big american tech firms i mean that's all he's railed against for the last decade european regulators regulating what he sees as his american company there are two things to keep in mind first in washington dc people don't look at the uk and they don't look at london the same way they look at brussels that's a strength for the united kingdom in this regard. Second, when a problem arrives and when the moment comes that it needs to be addressed, somebody always needs to go first.
12:53Everybody looks around. You sort of hope, oh, maybe you'll go first so I don't have to. Who's going to summon the courage? Look, at Microsoft, we're sort of playing the same role. We came out in 2023 and in 2024 and said, look, there's got to be a role for common sense rules here. No technology should be above the law. And I appreciate that the debate has taken a different turn over, say, the last 18 months. But look at 2026. When you look at AI agents that have escaped human control, when you look at models that are so powerful that they're putting the world's cybersecurity at risk, I would say, look, the moment has arrived.
13:37Let's speak calmly, maybe even softly, but firmly and start to bring people together. I hope that's the role we can play as a company. I hope that's the role the UK can play as a country.
13:50Will Bain: So there was this accord that came out of the White House, wasn't it? You were notably not one of the signatories as a company of that. But that was all about, I'm quoting the president again here, tremendous self-regulation. It sounds like from what you're saying to me without putting words in your mouth, you don't think self-regulation actually goes far enough. Then it should be one body somewhere that people are compelled, forced. What is it? Legally forced to put their models through to test. I would say two things. First, applaud. I applaud. We applaud as a company that statement. And we're happy to sign it.
14:21It's a good step. It's a necessary step. And fundamentally what it does is it says companies should have internal controls. It should have its internal auditors looking at things like AI safety. We've been doing that since almost a decade ago. And there ought to be some external evaluators as well, all to the good. It's necessary. I don't think it's sufficient any longer because the challenge for any self-regulatory regime is if it's left to the industry alone. Frankly, the good companies will do what you want. It's the companies that are not as responsible that you need to worry about. We should always worry that regulation can become too heavy-handed.
15:02It can slow innovation. And I think President Trump is right to remind us to think about even worry about those concerns. But right now, I'd argue that AI is moving forward faster than the speed of safety. That's not good for the protection of the public. It's not even good for the industry itself. And usually what happens, and you take, I'll just say, our Congress in the United States. If a product moves faster than the speed of safety, accidents happen. And then Congress moves faster than the speed of thought. And then you just end up with a dynamic where everything comes to a halt. I don't think that should be our goal either.
15:45AI is an extraordinary resource to grow the economy, to solve societal problems or at least help. Let's strike the right balance.
15:53Will Bain: The other side of regulation, of course, is what the punishments are. You breach it. So what should those be and who polices those? Well, I think all of these are questions to be discussed. I don't come here to London with all the answers. I don't think it would even be helpful if I thought I had all the answers. No, but it's important to start the conversation, isn't it, about what would work? Yeah, absolutely. And there are typically a variety of different regulatory approaches. The list is not that long. So it's like looking at a menu of a restaurant. Yeah, you figure out, do you want to go as far as requiring a company to get something certified?
16:32Licenses and things like that. Yes, you have licensing. You have pre-release review. You can have self-reporting requirements. You can have penalties if a company fails to comply with its obligations. Some of these things like incident reporting are more just disclosure obligations, again, as we have for privacy. So we're familiar with them across not just the tech sector but the entire business community. There is a lot that we can work with if we put our minds to it.
17:01Will Bain: When you're saying we're at that moment, I was just trying to glance down to tot up the number of quotes I've got from people kind of agreeing with you basically. Exactly, flagging the alarm. Sam Altman, Dario Amadei from Anthropic, even Antonio Guterres from the United Nations. Are you worried? I mean, you've had a great career in tech, what, three decades? It feels about as transformative in that period right as any period of technology. Do you sometimes look at it and think, oh, yeah, Dario Amadei is right. I am a bit scared at the pace of this. I do think we should worry more about the state of AI safety than any prior generation of digital technology, certainly in my 33 years in the industry.
17:46And certainly at Microsoft, we have made this a priority. We want to make it a priority. But I think there comes a day, and it's what I find when I talk with our employees, as I have been doing here in the UK. Look, everybody's got a family. Maybe their family is parents and siblings. Maybe they have children of their own. We should do this as a business, but more than that, we should do that for those that we care about. We should do it for the population. It's the right thing to do.
18:18Will Bain: So it's not hubristic. It's not hyperbolic. There is real threat here. I think we are best served today by treating this problem as something that requires a seriousness of purpose and real urgency. Because the truth is the technology has improved faster over the last one to two years than almost anyone expected. Don't wait another one or two years. Don't wait for a big accident to happen to finally take action. I think that the kind of action that's needed has become increasingly clear and the need for it has become apparent. And those concerns then spread out and they spread from being the ones that we've heard about, about the damage it could do to humans or the attempted hackings we're seeing of healthcare systems, all kinds that feeds into it feels like a public perception, particularly at home in the United States.
19:12But it's clearly an issue around the world as well about the public thinking, hang on a minute, this is going too fast for me actually as well.
19:20Will Bain: And that is being really manifested, particularly at home, isn't it, about people's opinions towards data centers that are the bedrock to all of this. How do you try and wrestle back that argument? Because right now, I mean, I'm looking at the polling down here, isn't it? 70 % of Americans who don't want a data center in their backyard, that's going to prove pretty problematic down the track. I mean, you need lots more of these and you're spending tens of billions of dollars on these. Perhaps it's because I've spent so many years working in this industry and I've spent literally a quarter of a century in the senior ranks at Microsoft.
19:56Whenever there is a public argument, whenever you find people who disagree with you, you fundamentally have three choices. You can go fight with them. You can run away from them. Or you can lean in and try to understand what people are worried about. And consistently, I have found that if you lean in and you understand what is bothering people, then you can go to work to solve the problem. And so there are two things that are really important with respect to data centers. The first is really around the world. If you tell people there's going to be a data center and you're going to be able to stream a movie, you're going to be able to share a photo, you're going to be able to do a video meeting or you're going to be able to use it and you're going to get better medical care in your community because of it.
20:49People go, OK, good. That's a good thing to build. If you say, no, it's for AI, then people go, oh, I'm not so sure because people have anxiety about AI. So you have to separate those two and we need to make sure that AI is being built in ways that people see benefits and feel safe.
21:09Will Bain: Do you think it is the one after the other? It isn't just the fact that these things are enormous. No, that's the second. No, you're absolutely – that's the second dimension. And the thing that is interesting about that is when you listen to people or you go, I love going to the communities where we have data centers or we're talking about building them. I mean, what I really like to do is you sit down with local businesses. They're often small businesses or people who are in nonprofits and they list what they're worried about. It's a very common sense list. I don't want this to increase the price I pay for electricity.
21:39I don't want to lose all the access to water I need. I don't want this to take away my jobs or more precisely, I get that this will create jobs. That sounds good. But I want to know, will my kids, my neighbors, my family have the opportunity to get some of those jobs? I want to know, is this going to just run up the cost of maintaining our highways and roads or is it going to add to the tax revenue of my community and make sure it doesn't make too much noise?
22:12Will Bain: And they are that at the moment, aren't they? But the truth is if we build them responsibly as we can, as we're really striving to do as a company, you can address every one of those problems. That's the set of commitments that we made as a company back in January. We spelled out pledges and we're going to work to implement them. We're putting in place people and budgets and signing up where communities want to legally binding commitments. That's the right way to do this in my opinion. Let's put a pin in jobs because I want to come back to that. I think I'll link to your own career and kind of hopes on that as well.
22:48Will Bain: The energy side of things, you're in a debate around one of your dead descendants in Virginia at the moment, or a series of dead descendants in Virginia at the moment on your round. Who pays for that power? Should it be you guys in the tech industry? Should it be the AI models, developers who are paying for the energy that goes into it? Well, almost always I would say – well, let me just say first. I'll tell you who it should not be. It should not be the neighbors. It should not be the consumers or citizens. So it should be the industry itself. And then you get into a question of, well, how do you allocate that responsibility?
23:22That's hard, isn't it?
23:23Will Bain: Because they just add it onto bills and then it's there. In a sense, yes. But the services are used so broadly that then you can spread those costs in a way that people – it's not the same as when you're asking a neighborhood or a local community to bear that cost. So our point is we need to pay our way. And that means if there's an improvement to a substation or a new substation that needs to be built, we should pay for it. If the grid needs to be improved, we should pay for it. If there's new electrical generation capacity that needs to be added, we should find a way to pay for it. It may be to pay for the investment or it may be to pay in an offtake agreement.
24:07So the utility provider says, yes, now I've got that covered. Now, you may get into situations where you should discuss then how does it get allocated between, say, the hyperscaler, the data center company, and a model company. Or you may find that you make all those improvements and then a big manufacturer wants to come to town and build a plant, which is a good thing. They want to use those improvements. And then that gets reallocated so that they pay for a portion as well. The most important principle I think though is clear, make sure that the neighbors, the community, the citizen, the consumer, that they're not seeing their electricity price increase because a data center has come to town.
24:52Will Bain: And last one on data centers, because we had Steve Hebb, boss of FTSE 100 accounting firm Sage, on this podcast a few weeks back. And Steve was talking about, I guess, whether you really need them everywhere, if you know, whether we can use the capacity that's there better. Let's, I think, I hope you'll be able to hear a little bit of Steve making his point here about how many we need in the UK. I don't really see why we need data centers all over the UK. They're just big sheds full of computers, to be honest. The UK does not have the capability to be a sovereign data AI nation. Just, I suppose it takes us back to your point about safety, is who has the ultimate switch off.
25:35Will Bain: But is that an answer, actually? That a lot of the computers, you guys in the industry call it, could come from somewhere else. That power could come from somewhere else, and we use it elsewhere in the world. Will that be somewhere that the technology will move down the track? It could. I would say two things. First, I think we should always respect the right and the ability of whether it's a country or community to make its own decision. If it wants something, OK. If it doesn't want something, OK. But then I also think that it always requires an informed choice. And so if somebody says, I don't want to see these built in the UK, I don't want us to see us have to invest in electricity for them, OK.
26:18But then it is important to understand, frankly, the downside of that. What does that really mean? It means that people's data in the UK is probably stored in a data center in Ireland or the Netherlands or Belgium or some other EU country. What does that mean? It means that that data is probably going to be governed by the laws of the European Union. It means that it will be more challenging for the British security agencies to protect the national security of the United Kingdom and they'll have to rely more on the security agencies and the like in these other countries. So life is full of choices and seldom is it, oh, 100 to nothing or zero or not to 100.
27:11And let people and countries decide, but let them decide with their eyes wide open.
27:16Will Bain: All of that is political, diplomatic almost as well, which feels like it seems like it must be more and more and more of your job. Well, it's been a long time since I've heard somebody say that the people of the United Kingdom should be subject to the laws of Ireland. But if that's what people want, it's not my place. It's not our place. Well, I think I suppose they're thinking about can we get around not having big data centers in the UK if they were using your power. I mean I think it's one of your customers actually. I think he uses Copilot to run their AI system but use it from your data center in the US or whatever.
27:48And the truth is it's even entirely sensible potentially for different customers in the United Kingdom have different points of view. I may be a business. I'm not worried about that. And I do business in these other countries. And then a home, a consumer, others. So let's have the conversation and let's understand that there may be room for differing views and different choices.
28:13Will Bain: Has your job become even more political? And has that changed a lot because of, I guess, how – I mean lots of politicians want to get involved in changes. But President Trump seems particularly fired up about AI and what AI means and about the fact that the United States needs to win in that. Has that shifted how your role has to be, I suppose, in relation with him? The truth is every year things seem to get a little more complicated in part because the technology keeps advancing. And as it advances, it's more powerful and also can be weaponized in new ways. It gets more complicated because, frankly, there's more divisions in the world today.
28:55Countries are less united. The good news is there's always room to work together in many areas. We basically have a simple philosophy as a company. Partner where we can, stand apart when we should. And that is a philosophy we apply to every country and every government. And I applaud President Trump for many of the things he is doing, frankly, to add to the electricity capacity of the United States, to encourage innovation and to put innovation to work, to attract foreign investment, all of these things. And then there are days that we disagree. And I remind folks, including in Washington, D.C., that Microsoft sued the Obama administration not once but four times all over the protection of privacy and the privacy rights of individuals including here in the U.K.
29:52and across Europe. And so the need really, it's an imperative. You've got to find a way to disagree without being disagreeable, without making things personal, with understanding that on any given day, it's almost like every relationship that matters. You agree on some things, you disagree on others and you find out how to maintain that relationship.
30:15Will Bain: I suspect you disagree with them around this. This is Vice President J.D. Vance speaking overnight in the United States about visas and your hiring processes at home. There has been no company in the United States, unfortunately, that has abused the system more than Microsoft. And let me just give you a couple of statistics on this. So in the last year, Microsoft laid off last year 6 ,000, 6 ,000 American workers. At the same time, the company benefited from 6 ,300 H-1B visas and almost 3 ,000 green cards. In other words, if you do the math, for every worker that Microsoft laid off, they replaced that worker with one and a half foreign indentured servants.
30:56Will Bain: You have servants in the United States, apparently. Not that I'm aware of. Let me say two things. With the same spirit, look, I think Vice President Vance is encouraging us and others to hire more Americans. And that's a good thing. It's something that we believe in and it's something that we actively do. Then I'll talk about where we disagree. And we said this literally within minutes after this statement was made. Whoever did the staff work for the vice president got the math wrong, got the numbers wrong. They got one part right. Yes, last year we did eliminate 6 ,000 jobs. Not all of them were Americans, by the way, but a large percentage, the largest by far were because most people we employ are Americans.
31:44But the rest of those numbers actually don't add up. He said we hired 6 ,000 foreigners to replace them. No, actually what we applied for was 6 ,000 visas, but 80 % of those visas were for people who already were working at Microsoft. They just needed a renewal or what's called a change in status of their visa. He said that we applied for 3 ,000 green cards. I don't know if the number is quite right, but what that misses is every single green card applicant was already a Microsoft employee. So those were not people who were replacing the people. Those were people mostly who were already employed.
32:29And not only that, it misses the fact that last year we also hired another 9 ,000 Americans to work at the company. We pay everybody the same wages. We give them the same benefits, the same stock compensation. We do not differentiate. It's not as if we pay foreign employees less than American employees, nor would we do the same thing, say, when we have an international workforce in the United Kingdom. And the other thing that I would say is when you look at the data, including of all of the companies that apply for visas, we are among the very highest in paying the highest wages. And that doesn't even account for the stock and bonuses we give people.
33:18So if the facts were right, I would join him and I would be among the first to say, wow, we need to think differently.
33:26Will Bain: Why are you in the firing line then, do you think? Because, I mean, that's a very personal attack. He's called you out by name as a company. There's thousands of tech companies in America who presumably use this visa system, thousands of companies full stop. There's you, there's Harvard, Yale. We know that the administration has had some beef with some of the Ivy League universities. Why is it you? Why are you in the firing line? I would love to – I will suspect I'll find out. But I will say this. Look, I've sat in meetings with Vice President Vance. He's smart. He's thoughtful. That doesn't mean that he – I or anyone agrees with him on everything.
33:59That's the nature of life. No one agrees with anyone on everything. I look forward to the opportunity to talk through the facts and I'll listen. I mean I think we need to recognize he has a message that says, OK, fine. I still want you to hire more Americans. We believe in hiring Americans in the United States the same way we believe in hiring the citizens of the United Kingdom here in the UK.
34:27Will Bain: Well, to that point, is it an opportunity for us here in the UK if this is going to be a prolonged – you're suspended from this visa system now. Is that an opportunity for us for you to hire more people here? A number of countries have, frankly, focused on that. UK, Ireland has done a good job. Canada has done a good job. It does point to a more complicated aspect of the world today. Let's just say if you want to build a global technology leader, either as a company or a country, you've got to have a world-leading technology workforce. And the basic recipe for doing that anywhere is almost always the same.
35:10Build out, at your core, people who are from that community or country. Invest in the education system so even more can graduate and come work for you. But add room for truly talented people from other countries as well who understand the world that our products need to serve. That's what we've long done here in the UK and in Ireland, actually. The two countries, both in different ways, have been hugely important and have been real international leaders. Look at London. Look at what we're doing for AI here in London with our Microsoft AI workforce. It's hire great people from Britain and add some from other places.
35:58Will Bain: Well, I'm going to go a bit parochial for our last couple of minutes, if that's OK, and talk very specifically about the UK. What helps us in that race then? Because that's clearly a challenge. We've made this decision politically. And that one's in the past of what we've done with Brexit. So we're clearly competing, I guess, somewhere across the sea, I suppose, in terms of Ireland as well. What would you be looking for? What makes Microsoft think of Britain as the place over Amsterdam, Dublin, etc? What do we need to be thinking about as a country? Well, let's be parochial. And I'll start with, hey, my wife and I lived here for the last four years before I joined Microsoft from 1989 to 93.
36:35We're here in London. Our son was born here. He's a dual national American and UK. And let's just talk about all of the things that I think make the UK great. It's got a strong workforce. It's got the rule of law. It's got the protection of human rights and a system that encourages and rewards creativity, including in the form of intellectual property. It has this extraordinary network of relationships literally around the world in the form of the commonwealth. And those relationships still matter and a common legal tradition and really common tradition of innovation all matters. And then you add to that, frankly, a lot of things that were not here as much when I lived here in, say, 1990.
37:28You've got a diversity of people, got a diversity of restaurants. You've got a place that people live.
37:36Will Bain: But that's a debate here too, exactly the same. We're talking about President Monson. That's a huge debate here, as you know, politically as well. Is that something you would, I guess, urge to think about, to resist? There's a much smaller economy in the United States in terms of visa rules or making it easy for access to that talent you were talking about? Ultimately, jobs go to places that are open to people and reward creativity and talent. And it's, again, one of these things where there's a balance that needs to be struck and to some degree there's a pendulum that is constantly moving a little bit back and forth.
38:13I don't think any country can ever afford to be a place that only creates opportunities for its own people and never lets anyone else in. Nor can any country afford to be a place that rewards and provides economic benefits to immigrants and leaves its own people behind. And you just take that. That has sort of defined the state of many countries and even our world over the last 30 years. Make – create more opportunity for your own people. Let your own people become more successful by connecting with the world as a whole. And, you know, location matters. It turns out that being an island off a continent is not a bad place to be.
38:57And when you are the island that is a bridge, not only across a channel but an ocean, that is an unparalleled asset and opportunity.
39:08Will Bain: And one that clearly is being explored by the current government again. Loop all these in together. You were laughing about this when we came up earlier when you were on with my colleague Sean a few years back. You had a lot to say about our regulatory system that made a lot of headlines around the world. And you talked about it being uninvestable. There's actually been a lot of change through successive governments of different colors around that regulatory system. Now that you come back now and you've been here, as you say, this week, does it feel like a place that that's making a difference, that it's easier?
39:38Will Bain: Because we're obsessed with growth here at the moment, aren't we? Do you feel there's an environment for that kind of growth here at the moment? Yes. I mean, like every country, there are challenges and frankly, every country needs more investment, domestic investment and foreign investment. Every country has opportunities to make regulation, if not lighter, at least faster. And I think the UK has been working towards it. There's opportunities for more speed and permitting, especially around electricity and the like. I would definitely call out in a very positive way, certainly the evolution of our relationship, I would argue the tech sector's relationship with the CMA, the Competition and Markets Authority.
40:26That was obviously the epicenter of the controversy that I found and we found ourselves embroiled in a few years ago. I actually think if you look at the world today, go back to this question of capability. Just as the AI Security Institute has this extraordinary capability to evaluate and test AI models, the CMA has a very professional, capable group of people who work there. I think Sarah Cardell, as the CEO and head of that agency, has really led it in a way that has helped it become more decisive, more open to just gathering more information to make fairer and more informed decisions. And interestingly, if you just look at this field of managing the economy, ensuring there's competition, there's a lot of tension right now between Washington and Brussels.
41:23that makes it harder for both Washington and Brussels to address many of these issues. And when you think about the future of AI on the economy, how it can spur growth, even how it impacts companies across the economy, how it impacts jobs, here too is a space where I think the UK has the capability and an opportunity to maybe not lead the world. That's a big, tall order. But by doing well. Dutch, by putting in place common sense rules, by ensuring, for example, that when you have organizations across the economy that are using AI, this is something that we've talked a lot about, that Satya Nadella, our CEO, has talked about, that they don't lose control of their own know-how, their learning, their IP.
42:12They don't see that leak over to model companies so that the model companies use that to train models to, frankly, outperform their customers. That would be bad for the economy and it would be bad for jobs in the UK. That's a good example of the kind of issue that the right kind of agency, including the CMA, can address. I also just think the business community as a whole can use its voice, both with its feet, so to speak, when it decides whose services to use or buy. But it's just the other imperative. We have to keep people safe. We have to keep the public safe as AI becomes more powerful. And we need to do everything we can to ensure that its benefits are distributed broadly across the economy in ways that actually create better jobs.
43:01Will Bain: We're right up on your time. I want to ask you one about those jobs to come up. But just a quick one on that. I've been struck in the kind of half an hour that we've been chatting. You seem very pro the UK, certainly in its position perhaps compared to Brussels in particular, which is obviously how we've kind of compared ourselves a lot, as you might imagine, in the last decade. I think, first of all, there's so many assets that the European Union has and each country is different. The UK can benefit from being a bridge at a time when the world is short of them, a bridge between countries. And by being a bridge between countries, it can do the world a lot of good.
43:41But perhaps more importantly for people who live here, it can do the people of the UK a lot of good by attracting more investment, by building trust, and by building trust in ways that will help the people who produce products here or services here bring those products and services to the rest of the world. So I would argue that this time of challenge is actually a moment of opportunity.
44:07Will Bain: And roll these two in together for me then in this last one because you talked about jobs. I'd like to know whether you're sort of very bluntly optimistic about what AI will do for jobs. And within that, I've got a lovely quote from you talking about, I think I'm right in saying if this anecdote is correct, that you insisted on having a personal computer in your first job back in the 80s. Yeah, in 86. That was one of your, what would be the thing that you would insist if you were starting your career again now that is a must, is a non-negotiable that would need to be in your office being a new person?
44:37I would say two things. Number one, first, I am more optimistic than many, but at the end of the day, it doesn't really matter. It doesn't matter whether you're optimistic or pessimistic. What does matter is whether you're determined. I think we should all be determined, frankly, to spend less time listening to predictions about what will be done to us. Let's spend more time together defining the future we want to create and working together to build it. Now, then let's talk about your second question. What is the future we should build? Let's build a future that enables people to use AI to do their jobs better, to make them more rewarding.
45:20That's what I pursued when I entered the workforce in the 80s and said, give me a computer on my desk. And the law firm that I was talking with said, first, what is wrong with you? We have people that will type up your penmanship. I said, I can write better. I can write faster. I can write more creatively. Give me that device. And so when people ask me, I'm in a university. What should I study? Or parents, what should my kids study? I always say, do three things. First, focus on what you want to do and what you want to be in life. Think a little bit about whether it could be automated away. But in general, the spark of human ambition and creativity will, in my opinion, persist.
46:06Never underestimate human ambition. If there's something that you love, pursue it. Second, learn how to use AI. Learn how to use AI to do whatever you do today even better in the future and in a way that gives you more job satisfaction. And third, remember, in a world with more technology, it's actually the soft skills that can be the big force multiplier and human differentiator, the ability to understand others better, to listen, to bring people together, to help people work together as a team. Oh, my goodness. That will always matter. Now it will matter even more. And frankly, look, look at the folks that you sometimes see in big jobs, the big boss in some of the tech companies in our industry.
47:00They're super smart. They're high IQ. Those soft skills, there's an opportunity for all of us to maybe outperform them.
47:08Will Bain: Those arts graduates everywhere. You better believe it. Brad Smith, the president of Microsoft. Really appreciate your time. Thank you.
47:18Will Bain: that was brad smith the president of microsoft speaking to us here on the big boss interview podcast if you don't already please do subscribe means you'll never miss an episode and while you're there if you do get a chance still trying to really grow that audience so if you could give us a five-star review and perhaps write something um nice about what you've enjoyed in one of the episodes for example or who'd like to hear from next that'll be massively uh one appreciated but two will help loads of other people find the podcast uh does help us boost us in those algorithms wherever you download yours from as well.
47:46Will Bain: So we'd really appreciate that. Thanks so much for listening to Big Boss Interview. Balance your trading strategy by adding futures. CME Group helps you manage risk and capture opportunities in all market environments. Capitalize on around-the-clock access to highly liquid global futures and options markets across all major asset classes. Visit your online broker and get started. See what adding futures can do for you at cmegroup.com forward slash podcast. Derivatives are not suitable for all investors and involve the risk of losing more than the amount originally deposited and any profit you might have made.
48:18Will Bain: This is not a recommendation or offer to buy, sell, or attain any specific investment or service. Innovation has always kept the world moving forward, but it also comes with risk. As technology and AI evolve, AXA XL is facing into risk. As a leading global insurer, we constantly push the boundaries to protect your data and network from cyber risks. AXA XL, facing into risk for a future to be imagined. This advertisement doesn't constitute an offer or solicitation nor a description of any products or services of AXA XL.
Artificial intelligence is advancing faster than the safeguards around it, and industry self-regulation is "not sufficient any longer". That is according to Brad Smith, president of Microsoft. Speaking to Will Bain on BBC's Big Boss Interview, Mr Smith said the world should worry more about AI safety than about any previous generation of digital technology in his 33 years in the industry.
Microsoft is one of the world's largest AI developers and an investor in OpenAI. Mr Smith's comments set out a position that goes beyond the "tremendous self-regulation" praised in a recent White House AI accord.
"Right now I'd argue that AI is moving forward faster than the speed of safety," Mr Smith said. "That's not good for the protection of the public. It's not even good for the industry itself."
Mr Smith set out three measures. The first is a "safety brake or emergency brake" on any AI system controlling autonomous devices or critical infrastructure, an idea often described as a kill switch. He said it should sit in the companies running the models and in the cloud infrastructure they run on. The second is a legal obligation to report incidents: if AI agents "break loose" and break into systems, the victims should be told, as they are after a data breach.
The third measure, which he called the biggest and most important, is whether government evaluators should have to sign off frontier models before release. "Just remember, right now, your washing machine is subject to more safety rules than AI," he said. He added that cartons of milk and passenger aircraft are subject to inspection and certification.
He said the UK was the country best placed to lead. The AI Security Institute, he said, has the strongest technical capability in the world to test AI models: "It's the United Kingdom. They are the best." In his account, the US began building similar capability and then slowed down. Asked about Anthropic declining UK testing of one of its models, he said the question now was when developers should be required to satisfy a government regulator that a product is safe before it is released. He described Britain as "a bridge across the Atlantic and across the English Channel". He also noted that in Washington, people "don't look at London the same way they look at Brussels".
Mr Smith also responded to Vice President JD Vance, who had accused Microsoft of abusing the H-1B visa system more than any other US company. Mr Smith said the vice president's staff "got the maths wrong". He said about 80% of the company's 6,000 visa applications were for existing employees, every green card applicant was already employed by Microsoft, and the company hired 9,000 Americans last year.
On data centres, he said the cost of new grid capacity "should not be the neighbours… it should be the industry itself".
Presenter: Will Bain Producer: Olie D'Albertanson Editor: Henry Jones