AT&T CEO John Stankey on AI, Return-to-Office Mandates, and Surveillance Law |  | Courtesy of AT&T |
|
|
| | BY HARRY BOOTH
Reporter, TIME |
|
| | | When Alexander Graham Bell made the first telephone call in 1876, few could have predicted what would grow from that moment: A company that would invent the transistor, the solar cell, and eventually become what is today America's largest telecommunications provider. AT&T celebrated the 150th anniversary of Bell's first phone call on March 10. As it marked the anniversary, the company announced an investment and spend of more than $250 billion to build high-speed networks and support "the future of U.S. advanced connectivity." | The anniversary comes as the company faces another technological transformation. Its CEO John Stankey, who has spent his entire career at the company, says any executive who thinks artificial intelligence won't disrupt their business is "naive." | Meanwhile, the future of Section 702 of the Foreign Intelligence Surveillance Act, which determines how the government can compel companies like AT&T to assist with surveillance, will be determined in April. Stankey tells TIME the law needs reform, arguing that the focus on telecom companies is outdated, as much of the data on AT&T's network is now encrypted. "We need to bring that discussion well beyond the telecom companies into those that run the apps, the operating systems, and are part of the broader ecosystem," he says. | This conversation has been condensed and edited for clarity. | AT&T invented the transistor, the solar cell, Unix, and information theory. As AT&T celebrates its 150th anniversary, where do you see this chapter fitting in? | We've moved from the era of the phone company 150 years ago to the era of a networking company today. My belief is that the next chapter is about being the foundation for interconnection and communication in the AI generation. All this infrastructure that's being built to allow for great inference and increasing productivity and people's knowledge: it's got to be connected to the place where the inference is required, to the work process. That's our role. It's that fabric of connecting that we'll continue to carry forward, and that I'm excited about. | You've been at this company your entire career, essentially. In that time, you've seen the technology reinvent itself time and time again. Now we're in another moment of upheaval with AI. How do you think about navigating the pace of change? | Since you brought up the fact that I've been here my entire career, it's an exercise to go look at how little revenue is generated on any product that existed when I started working for the company today—it's sub-25% of our revenues. So the company has been accustomed, because we're a foundation-of-technology business, to reinvent itself, to innovate on top of those changes. I think we've done a good job of embracing that over time and accepting that the business model has to evolve. | We evolved at the dawn of the internet in '99 and 2000, and really became a data networking company. That was a fundamental shift in how we built networks, how we allocated capital, how we built products. We're going to see another one of those moments. I believe anybody running a business today who walks into this moment thinking they're not going to be disrupted or impacted by AI is being a little naive. You have to embrace it. You have to understand that there are going to be great things that come with it that you have to figure out how to go with, and at the same time disassociate and disconnect yourself from those parts of the business that will fade away and no longer be relevant. We'll have those in our business as well. That's why we're doing what we're doing, transforming our company, designating parts of our business as legacy, reporting it as such, walking away from it, and leaning into what can be the future. | You lived through the telecom bubble of the late '90s, when the industry overbuilt on the promise of internet demand that took years to arrive. A lot of people are asking whether the scramble to build AI infrastructure looks similar right now. What do you see that's the same and what's different? | There was a degree of overbuilding during the telecom bubble, but most of that was in long-haul fiber. And there's a difference in this industry between what's done in the core of a backbone and what's done to get to end users, customers and businesses. There has not been an overbuilding of technology at the edge. In fact, that's where we're focusing and dedicating much of our investment, because those users are going to need to get onto those robust backbones. So I don't worry, from a telecom perspective, that there's going to be over-exuberance. Is it possible that in other parts of infrastructure there may be money that gets stranded in some places? That's possible. Whenever there's a significant TAM [total addressable market] to address, large amounts of capital go in and it's not all going to be successful. That's the reality of a great opportunity like AI. | I'm asking specifically about data centers. | If data centers were overbuilt, I don't think our efforts to serve them are going to end up being consequential or material if they become stranded or less than productive for some period of time. I do worry about the people building the data centers and how much money they're putting into them. It could be impactful on their business model. But the way we build networks today and how we deploy fiber, the marginal cost is dramatically different than it was in the '90s, and I think we know how to manage that going forward. | In the memo you addressed to staff in August, you take a pretty hard-line stance on returning to the office. There's a body of research suggesting that hybrid work is the sweet spot. What convinced you that getting people back in-person was the right call? | I wouldn't characterize it as hard-line. I took a definitive stance on what I thought was necessary for our business, given where we are in our transformation and life cycle. Every business has a business model and a time in their maturity where different strategies are appropriate. For our company right now, in this significant time of transformation, especially coming out of COVID, where for several years we were unable to do the things with our employee body that are really important: building a fabric, mentoring, allowing people to establish networks, having informal networks to learn about their discipline and expertise, because COVID and working from home sanitized a lot of that out of the daily work routine. Given the fact that we are now investing in significant projects—about seven of them, multi-billion-dollar projects that will go on for multiple years, all requiring high interdisciplinary execution—for our company right now to pull that off, we need to be shoulder to shoulder, working dynamically every day. That's just what's right for us in this season. | The other thing I'm well aware of is that societally, we have now started to underestimate the importance of daily personal contact with people. Strictly leaning on digital means of interacting allows people to distance themselves, to be more clinical, to not understand the plight of individuals, their particular needs and circumstances. Those things are better managed when you can see people eye to eye, break bread with them, understand the dynamics going on, and build the kind of culture that withstands the divisiveness going on in broader society. I feel this is right for our business right now. | Section 702 of the Foreign Intelligence Surveillance Act, the law that governs how the government can compel telecom companies to assist with surveillance, is being renegotiated right now. You run the country's largest communications network—the government literally cannot do this kind of surveillance without companies like yours. Do you think Section 702 should be reauthorized as-is, reformed, or perhaps allowed to lapse? | My personal belief is it needs to be reformed. In any civil society, in a trusted democracy, there should be a belief that government will handle information in a way that's appropriate and in the best interest of its citizenry. But this is less an issue of telecom reform—it's more of an issue of tech reform. Today, much of the traffic that flows over our network is encrypted, and it's not encrypted by us. It's encrypted by those who author the applications or run the operating systems which devices operate on. To think that we're going to manage the new wave of what is appropriate, legal and prudent law enforcement—to keep society safe, ensure children are safe, ensure that people aren't conducting deviant behaviour—we really need to think about this broadly. We need to bring that discussion well beyond the telecom companies, into those that run the apps, the operating systems, and are part of the broader ecosystem. I do believe it's time that we have a more holistic policy on that. | Thinking about what that would look like, are you talking about whether encryption should be regulated? | It has to start with the foundations of what we've used for many years: what is the bar that somebody who believes they should be able to get access to information needs to jump over to receive a compelling order to produce that information? There are clearly some revisions that need to go on, probably some improvements and checks and balances that need to occur, and I think there's a broader application of where that needs to happen. Getting industry together with those in government who care about these things, and trying to understand how we negotiate our way to something that everybody thinks is a step forward—that's probably what needs to occur. And I can probably project that in that process, not everybody's going to feel good about the outcome. | | So what you're saying is that right now, with encrypted telecommunications, there's no bar the government can clear to gain access, and that's a problem? | I think the counterparts of mine in government would suggest the process today isn't moving as cleanly, swiftly or effectively as they would like. Whether or not that's the case, that's a broader discussion that a lot of people need to engage in. And since we're not the ones doing the encryption, I don't consider myself necessarily to be the expert on what the right way is to manage or handle that. | The Lumen deal just closed. What does AT&T look like in five years that it couldn't have without it? | If you go back five years ago, when I walked into this job, I said this company needs to be an advanced networking company built on the foundation of the most extensive fiber network in the United States, offering the best connectivity across a variety of technologies including a nationwide wireless network. The reality was, five years ago, while we had the largest fiber network in the United States, we weren't everywhere. My belief is that as we've talked about with the evolution of AI—picking up high-performance workloads to enable autonomy, robotics, the right kind of inference at a point of need—it's going to require high-bandwidth two-way networks. In particular, really high-bandwidth upstream, because to get into the cloud and into these LLMs, you're going to have really high demands on performance: lower latency, higher throughput and bandwidth, more ubiquitous connectivity everywhere. | Lumen was about extending our fiber footprint in places where historically we have not invested aggressively. We picked up great fiber infrastructure in metropolitan areas and markets where we didn't have that kind of presence. This is just part of a march for us to get to a point where we can say: anywhere in the United States, we have a network that can pick up workloads, whether someone plugs into a wall, gets on wireless, or runs a business. All those workloads ultimately have to land on fiber. Lumen is the accelerant in those places where we weren't as dense, and gets us another step closer to the most pervasive and extensive nationwide fiber network in the U.S. | | Share the Leadership Brief by clicking here. |
|
| |
|
Commentaires
Enregistrer un commentaire
Thank you to leave a comment on my site