Managing the Future of Work

Coursera's Greg Hart: From MOOCs to skills for the AI economy

Managing the Future of Work Podcast Avatar

The online learning giant aims to be a general-purpose skills engine on a global scale. President & CEO Greg Hart explains the delivery model, from ramping up its enterprise business and expanding its credentials portfolio to employing generative AI without sacrificing the human element.

Transcript

Greg Hart: “AI can be phenomenal at helping somebody practice something but can be bad for retention. So it can be good for completing tasks, but bad for actually retaining knowledge of how to do something and actually really having the skill. I think that’s going to be one of the core challenges. How do you make sure that you’re using this phenomenally powerful tool in a way that doesn’t just help you get something done, but actually helps advance your capabilities as an individual?”

Bill Kerr: There’s an obvious appeal to online learning as the answer to a rapidly changing labor market, but access to courses is only one part of the problem. Workers need time and support to learn, opportunities to practice, and employers that can recognize and use those skills. As AI and demographic change reshape work, the harder questions are how learning connects to jobs and how mastery is assessed.

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Bill Kerr. My guest today is Greg Hart, CEO of Coursera. Before joining Coursera, Greg had a long career at Amazon where he led Prime Video globally and the Amazon Echo business. Coursera recently combined with Udemy bringing together Coursera’s individual learner businesses and university partnerships with Udemy’s enterprise business and instructor marketplace. We’ll talk about what that combination means for how workers build skills and employers access talent. We’ll discuss what it means for an employer’s business needs to become a learning path and how AI is altering learning and work. We’ll also consider whether credentials and skills records can show what people can actually do. Greg, welcome to the podcast.

Hart: Thank you very much, Bill. Great to be here.

Kerr: Greg, we mentioned a little bit of your background at Amazon, but talk about your path to Coursera.

Hart: Happy to do so. So, started at Amazon way back in 1997, spent the first 12 or so years in Amazon’s retail businesses—so it’s e-commerce businesses in the media space in the US and in Europe—then had the great good fortune to be able to serve as Jeff [Bezos’s] technical adviser, which is sort of a chief of staff role, for two years. Jeff, during the course of that, created an idea that he shared with me and a couple other people: two sentences. He then asked me to take those two sentences and turn them into a product, and that product became Alexa and Echo. So led the launch and creation of Alexa and Echo. And then my final chapter at Amazon was leading Amazon’s global Prime Video business and technology. And then spent not quite four years at Compass, an online real estate brokerage and platform, first as chief product officer and then as chief operating officer, and then joined Coursera in February of 2025. And what drew me to Coursera was the combination of the mission—the ability to have a real impact, helping people gain the skills that they need to be successful in a rapidly changing world— and the opportunity from a business perspective that I saw with Coursera. And I also felt a real strong appeal. My father’s a doctor; my mother is a school psychologist. They had both dedicated their lives and their careers to helping people. My career had been about scaling complex technology to deliver better customer solutions, but there hadn’t been the same sort of “social good” aspect. And so that part of Coursera’s mission certainly appealed to me, greatly.

Kerr: That’s so cool, and I mean, just to begin with, that there’s a household product that is in half of America that you helped put your touch on. And then, also, think about bringing the Amazon world of scale and digital intimacy and all that they know into this environment. Do you categorize a Coursera as similar kind of connection to the social impact and mission and kind of something beyond just the business side of the work as being a cultural theme there?

Hart: I definitely think so. I mean, to me, the business of Coursera and the mission of Coursera are deeply linked. And so, if you go back to the founding of Coursera, back in 2012, Andrew Ng and Daphne Koller co-founded Coursera, and it grew out of a real view of leveraging technology to enable anyone, anywhere in the world, to have access to world-class learning, and by giving them that access, to enable them to transform their life. But that mission is really only possible to the degree that we’re able to be successful as a business. We are a for-profit institution, and it’s really critical that we continue to grow the business and improve the business so that our mission and the impact of that mission can scale.

Kerr: Yeah. I’m sure most every listener has heard of Coursera, but perhaps we can take a moment to just understand what’s becoming an ever more intertwined, complex, multifaceted kind of education workforce ecosystem. What’s the distinctive things that Coursera brings that colleges or apprenticeships or other workforce programs aren’t doing? How do you integrate yourself, and as you think about that business-side edge, where’s the place that you want to most stand out?

Hart: Well, Coursera has a really rich global ecosystem that starts with all the amazing content that comes from, first and foremost, some of the leading universities and educational institutions in the world. So, obviously, Andrew and Daphne were professors at Stanford, and then of course, we very quickly expanded, now 14 years ago, to encompass content from Yale, Michigan, Penn, Duke, Virginia, Oxford, Cambridge, et cetera—the leading educational institutions in the world—and the leading technology companies in the world—content from Google, Microsoft, Meta, Anthropic, Amazon and AWS, IBM, Adobe, et cetera. DeepLearning.AI, Andrew’s company. And then the incredible fabric of learners that we have around the world who leverage the platform, and companies that benefit from and leverage Coursera to help upskill and reskill their organization. So we like to think of it as very much a virtuous circle of learners, organizations, and then content creators that together create a really rich, self-reinforcing platform that has phenomenal information about how people learn and about what is most effective at learning, and then amazing content that helps people gain, and organizations gain, the right skills.

Kerr: And continuing on that, just sometimes you hear companies declare themselves to be a B2B—business-to-business—B2C—business-to-consumer—maybe B2B2C. Is there a mentality inside Coursera that fits into one of those buckets? Because you’re obviously focused on the individual learner, but you mentioned there the importance of the business that is looking to go through a transformation. You have the universities as part of it.

Hart: So Coursera’s business had historically been two-thirds consumer, one-third enterprise. The thing that we are selling is learning that is created by some of the best human experts in the world. So we are leveraging content and learning experiences that come from those universities and industry partners to do that. So we also have a multifaceted relationship, because oftentimes we will both serve as a learning platform that an institution might be using, but we also might have content coming from that same institution that is available across the platform to deliver that learning experience either to individuals or to other organizations. And so, when we think about our B2B piece of our business, that is really across three different segments: working with actual businesses—so the core enterprise offering; working with educational institutions to augment their offering to their students—and so we call that Coursera for Campus, where a university, an educational institution, will leverage Coursera to augment their curriculum and make that available to students and oftentimes to faculty and sometimes to alumni as well; and then we also work and support governments at the complete federal level as well, depending on the nature of the relationship with a given government around the world.

Kerr: And as we mentioned a little bit in the introduction—I’d love to spend a bit more time on it—part of that coverage across so many spaces comes through the merger, and Coursera’s history was more connected to the individual learners, the universities; Udemy was more maybe the employers and the instructor side. Was it simply those kinds of interconnecting pieces, or was it scale? What are the things that drove that merger? And these things always take time to integrate and really see the fruition of the vision that was at the start, so tell us a little bit, Greg, about how you’ve been working to bring those pieces in alignment.

Hart: So Coursera, as I mentioned, was two-thirds consumer, one-third enterprise. Udemy is the inverse of that. And the businesses are roughly the same size. And so now that we’ve combined, we have a business that is 50/50, enterprise and consumer. And the interesting thing about the businesses is both companies historically have had very similar missions and have gone about them in relatively similar ways, but with important differences. So Coursera’s growth came out of, originally, higher education and then partnerships with content creators across the leading technical organizations in the world—the tech titans, if you will. Udemy, on the other hand, has this amazing subject matter expert network of instructors from around the world who are creating content at a really rapid clip that is very tied to what is happening in work at this moment. Because both platforms and both companies were investing in very similar types of features, it would enable us to actually make faster progress delivering features for our customers, whether those are individual learners or for enterprises. And that is what we’ve embarked on now. What we’re calling Project Helix is the AI-native, new platform that we are creating to do an even better job of helping organizations identify the critical skills that they need in their workforce to then create personalized learning paths for every single employee within those organizations, to not just deliver courses, but to deliver a verified mastery of skills. People need to continuously be upskilling, and that is what we are building Project Helix around: an ability to have very personalized learning paths that deliver skills in the moment and in the flow of work and to have those verified—assessed mastery of a skill that can be applied in a workforce setting. And then to ideally link that to really measurable business outcomes so that organizations are gaining the benefit of that. And then to do the same for actual individual learners on the consumer side of the business.

Kerr: There’s so much there. It’s going to take us a little while to unpack, I think, all those different elements. So let’s start with the business goals side. So you’re working with a company—a midsize company, a larger company—and talk us through how you are going to define those goals and set out this path. And is it where the company says it wants to be? Is there a reality check about labor market demands and where the... You can imagine this being too ambitious, not ambitious enough. Or how do the goals get then defined into the skill parts of the equation?

Hart: So what we really try and do is start with the customer and really just deeply understand what are their business objectives that they’re trying to accomplish and how do they view the skills, or the gaps in skills, that they have within the organization as things that they need to address and where we can partner with them. We have invested, over the years in building out a career and skill graph that sort of maps out skills across the world; it’s based on our own internal data, also based on data from third parties, externally, that we take in, and it basically looks at the different professions in the world, the roles within a given profession, the levels within a given role, and the skills needed for a given level in a given role in a given profession, and then maps that to the content that we have that helps build and deliver those skills. And so, when we work with an organization, with an enterprise, a business, or a government or university to understand what they’re trying to accomplish for their constituents—whether they’re employees or citizens or students—we then say, “Okay, well, if you’re trying to accomplish these specific goals, here’s the customized pathways that we can create with you to enable that.” And one of the things that we’re very much focused on with Project Helix is to have that be far more interactive and personalized, down to the individual employee level, because even if you and I are in the same role at the same level doing the same job, we each have different past experience.

Kerr: I want to come to the individual in a second, but let me stay one bit longer on the business side. How long does that typically take? So imagine a company is just starting out, they contact Coursera. What’s the time to build around just this skills perspective inside an organization? And one of my reasons for asking that is I was just on a panel about an hour ago, where the panelists were all talking about wanting to have skills inside the business, wanting to do it, and yet it was pretty clear across the panel just how amorphous that concept really is and how difficult it is to actually get down to “What are the skills that we need right now, what are we going to need in a few years, and how do we integrate a goal perspective across that?”

Hart: We can move at whatever pace is needed, but oftentimes—and I understand this as a CEO—oftentimes an organization may not have perfect clarity on exactly what they need to accomplish. They have a general business objective that they’re trying to accomplish. Okay, well, how do you actually then take that and bring that down to the department level? And so really understanding, “Okay, well, what is the current skill level and what are the current tools that a given department is using?” is typically a part of that conversation. We build tools, though, that make it very easy for managers or for admins in an organization to do that. So, on the Coursera side, we have Course Builder, which is an AI-driven customized learning pathway that enables a manager within a given department or an administrator within a company to leverage content that they have. Because often organizations, they don’t just want to leverage the amazing content that we have, they also want to combine it with things that they have internally, that are relevant to their company, to the context of their industry, to a given department.

Kerr: So that’s very helpful, because it helps me to also internalize it—not necessarily taking over the entire skills of the organization, like soup to nuts. It may be that you’re starting with a few specific interventions of skills that you want to have, and maybe there’s a compliance part of being in pharmaceuticals that they’ve already got a skills platform that’s delivering against that, and Coursera’s coming and helping them with the AI functionality that sits in that skill base as well for doing that.

Hart: Well, I would say both the AI functionality, but also increasingly a focus on human skills as well. So as AI is able to do more and more of the technical part of any given role and able to automate much more of what had previously been done manually, I think there’s an increasing focus on the importance of human skills. So leadership, communication, flexibility, creative thinking, teamwork, those types of concepts. And we’ve seen a real increase in demand for those, both for individual learners, but also within enterprises as well, who are really recognizing that, to get the most benefit out of this new technology, AI, that is transforming everything, people actually need to continue to develop their human skills as well as their skills around AI.

Kerr: Let’s continue on the individual side now. What I loved about the Project Helix approach is that it’s combining both where the organization wants to go with personalized, individual-level skill development that’s in a consistent path of that organization, which we think is, of course, best-in-class for the workforce transforming alongside the business. But sometimes, you have that “Maybe my ambitions as an individual are different.” How do you grab that individual-level perspective about what they want to accomplish and work it in?

Hart: Well, one of the things that we’re very focused on with Project Helix is really building around that exact dynamic. And so Project Helix has a more conversational interface to make it much easier for us to do a better job of understanding what the learner is trying to accomplish, both from a long-term career objective perspective, but also from their point of view on the next step in that journey. And sometimes that point of view is going to be really well-informed. Many times, though, I would say that is not the case. People have, again, more of a high-level, like, “I want to build a career. I really am interested in data. I want to build a career as a data analyst,” but they don’t necessarily know how to go about doing that. And so Project Helix is envisioning an experience that is much more of a conversational back-and-forth that helps us better understand the learner’s objective, but also helps inform them about some of the considerations for that and some of the next steps, and then helps assess where they are in that journey today, assessing that to then build a more personalized learning pathway that helps them achieve that objective.

Kerr: Do you get in settings where the payee may then differ? So if it’s a business-focused path stream, the company is supporting the training that I’m taking, but maybe there’s also an opportunity for me to say, “I want to go in this other arc.” It could even lead me out of the organization. And at that point, it becomes something that’s more self-paid in order to deliver against that goal.

Hart: Yes. I mean, historically, we’ve seen organizations use Coursera and Udemy in two different ways, sometimes both at the same time. One will be a general learning and development benefit that is provided to employees, and so then it is more self-guided. But the other one is much more about, “Well, there are specific skills that we want our employees to gain in these areas.” One of the things that I have a strong belief in is that the enterprise and consumer sides of our business inform one another. And so, on the enterprise side of the business, we get a really good signal from the companies that we support and that we work with on the objectives that they are trying to accomplish and the skills that they want their employees to gain, by department and by role. That is a fantastic signal that we can apply on the consumer side of the business. As individuals are interested in exploring careers or next steps in a given profession, we can leverage all of the data that we have on the enterprise side of the business—of course, in an anonymized, aggregated way—to help inform the right next steps for an individual learner who has an interest in a given area. And so we do see a lot of symbiosis across the two, and of course, they’re all leveraging the same platform.

Kerr: Yeah, it goes back to just the many different parts of the ecosystem that you observe and your position from there. You mentioned at the beginning of the Project Helix description that it wasn’t just about broadcasting the training out, but that there is a way to observe whether someone could actually use the skill at work. And I’d love for you to expand a little bit about how that gets accomplished and what maybe even credentials or other sorts of things get measured in that process.

Hart: We’ve spent a lot of time on building the right assessment tools. And that’s an important part of the certificates and the micro-credentials that people earn on Coursera today. We’ve seen enrollments in micro-credentials grow 33 percent, year over year. We’ve expanded our catalog across our professional certificates and specializations by 90 percent, year over year. So we now have more than 3,000 different programs there. And one of the reasons for that is that micro-credentials are increasingly being recognized by every single part of the ecosystem as being really critical and having a big impact in driving real outcome. So 87 percent of graduates with micro-credentials report securing a role that’s aligned to their field within 12 months of gaining the micro-credential. And that is an incredibly powerful stat in a job market that is very challenging, particularly for recent graduates. And then we see that 94 percent of employers are willing to offer higher starting salaries to graduates that have micro-credentials. So for an employer, obviously that real-world benefit is a workforce that is more capable of accomplishing what they need to, to help drive the business forward. For an individual, it’s enabling them to do a better job of gaining those jobs. And we see that reflected in those stats. We also see that universities are using micro-credentials to close gaps in their curriculum. And so one of the Coursera for Campus use cases is a university that leverages Coursera to augment their curriculum and their offering. So a great example: in Vietnam, we work with FPT, which has a network of different universities, and they have leveraged Coursera to expand their reach and ease faculty workload to enable them to serve more students more effectively. And so 20 percent of all academic credits at FPT are now earned via Coursera, and that has been one of the factors that has enabled them to increase their student population 44 percent in a single year, which obviously they could not do, just because of physical facilities and faculty hiring, if they were just relying on in-country hiring, within Vietnam. So lots of great examples there. As we shift to Helix and developing Helix, we really want to build on top of that and build something that does a really good job of assessing mastery in a very practical way. So, with role-plays, with labs, with applied portfolio reviews, with things that really reflect the type of work that’s being done in the workforce so that people can have a skill badge they can store in a skill wallet, a skill passport, a record to demonstrate to employers that they have real-world mastery of skills that matter.

Kerr: And how far are we along the path toward that sort of credentials wallet or portfolio that could combine things from Coursera maybe with other micro-credentials? You got great examples of the wage pay and premium that come with it. But just at the broad labor market level, are we early innings, mid-innings, late innings?

Hart: Well, we’re still in the building stages for Project Helix. I would say that the proof will come as we are able to demonstrate to employers that those skill assessments actually really do demonstrate mastery. And I think part of that is just the recognition that everybody has—not just employers, but also that universities have as well—that the relevance of a degree that was earned years in the past decays at a faster and faster rate that is inversely related to the pace of technological change. So as technology speeds up and things get created at a faster and faster rate with AI, the decay rate also speeds up. And so, we think that it’s really essential that people become lifelong learners and continuous learners.

Kerr: There’s so much innovation that’s being contemplated around how we price or try to claim some of the value that gets created, especially in digital places. And you can imagine, given the scope of Project Helix, the way it touches workers, and the value, that maybe you would... Coursera would be signing up to say, “We are taking responsibility and would anticipate being compensated based upon mastery, or skill use, or the mobility of the work, or the wages, or just other more outcome-based maybe pricing structures.” Is that something on the horizon, or is it still more traditional number of seats and use cases is the likely state of play?

Hart: Right now, most businesses in our space still are more in that traditional seat license or ELA-type approach. I think the shift over time will be to something that is more outcome related. We’re going to have one perspective on it, based on all the information that we see from usage of the platform, assessment of skills, delivery of mastery of those skills, et cetera. The organization, the enterprise, is going to have another because they’re going to be looking at output metrics, effectively, that come from that skill benefit. So they’re going to be looking at productivity and quality of work, et cetera. We are not necessarily going to have that level of visibility into that. So we want to focus on what we can measure, which is skill delivery, skill mastery. And then we want to make sure that we’re doing that in a way that then translates to output benefits that organizations really value.

Kerr: I’ve spent a little bit of time, not surprising, on AI. And when the LLMs and gen AI took off, Coursera was kind enough to be a part of one of our very first case studies of trying to understand the impact that it could have, and thank you for that. And I want to just continue with how is Coursera using AI in its own sort of work and the stuff that you’re doing internally, and just overall function of AI inside your organization.

Hart: Very shortly after I joined as CEO, I felt like we weren’t focused enough internally on AI usage and diffusion of AI skills within the company as we needed to be. So one of the things that we did is, we started a series that we call AI Sparks. And this was in summer of 2025. And the goal was to enable everyone to have access to different AI tools and to really have experimentation as the goal. The only way you get to proficiency is by enabling people to actually spend time using the tools, understanding what they’re good at, and particularly, given the pace of change of those tools, trying to keep up with that. And so, the bottom-up effort through AI Sparks was providing every department within the company with budget to use access to ChatGPT and other tools, Gemini and others, to really experiment and see what works and what doesn’t, and then share that organically. So every month, we have an AI Sparks meeting—that’s probably the best attended meeting that we have within the company, where people volunteer to share how they’re using AI within their work. That’s been fantastic in terms of sharing through osmosis what’s working at an organic level. But that’s not sufficient. We also wanted to have a more centralized model that enabled departments—not just individual employees, but departments—to actually create tools that better leveraged AI. And so we created a team. Well, we had a team, it’s based in India, that was a services and operations center [SOC]. What we really did is we decided to turn that from a smaller team into a somewhat larger team that really worked at a much faster clip. And so any department in the company can submit an example of a workflow that they have been struggling to try and automate and to leverage AI to automate. The SOC will very quickly turn around, within one to two weeks, a pilot of that, and then they’ll pass ownership of it back to the department. But they won’t just give them a tool. They will also train them, not just in how to use the tool, but how to keep improving the tool. We’ve seen a 95-plus percent success rate of the projects that the SOC tackles.

Kerr: I’m curious whether you have, not just maybe today, but longer-term visions or perspective as to whether AI is going to improve how durable, how long-term learning is. So one thing is that AI can maybe make it more entertaining in a way that keeps people engaged, which is great by itself. I don’t want to discount that. Another thing is that maybe AI helps us personalize the learnings a little bit better, which is great. But also, as we think about just sustaining this and keeping people curious and involved and so forth, do you think that there’s something we’ll be able to unlock in that way?

Hart: Well, I think it’s a really great question, because one of the things that we’re starting to see increasingly get commented on and written about is that AI can be phenomenal at helping somebody practice something but can be bad for retention. So it can be good for completing tasks, but bad for actually retaining knowledge of how to do something and actually really having the skill. And this is something that Andrew Ng, our co-founder, has commented on in the past and that we’ve seen studies from Wharton and from others. I think that’s going to be one of the core challenges, literally, for all of humanity. It’s not just for Coursera, it’s for every person, is how do you make sure that you’re using this phenomenally powerful tool in a way that doesn’t just help you get something done, but actually helps advance your capabilities as an individual? And I think AI can play a role in that, actually, if you structure it in the right way. And one of the things that we’re trying to do with Helix is to build a system that is minimally helpful, not maximally helpful, because AI is going to be as helpful as it possibly can be if you just ask as an individual to Chat or to Claude or to Gemini, et cetera. That’s its role, effectively. But there are ways that you can actually train the system and structure the system to provide the right level of assistance so that it’s actually helping you with retention, if it’s structured in the right way. You have to bring pedagogy to that, you have to bring structure and rigor to that, and you have to bring reinforcement to that. And so, having learners not just whiz through something, but continuously revisiting that and testing them on that so that they’re gaining, through repetition and through reinforcement, true mastery of a skill and knowledge of how to go about doing something. And ideally, that it’s not just a task, but it’s actually an understanding that broadens their horizons and helps them understand connections between things that they might not previously have seen. Because that’s what the best professors do. They help you to connect the dots in the learning experience in brand new ways and effectively forge new neural connections. We believe that AI can help do that if it’s done in the right way, and that there is a very strong human presence within that. And that’s where our grounding, coming out of Stanford and our industry partnerships and our Udemy practitioner network, is so important.

Kerr: That’s great. One thing we haven’t touched upon so far is that you’re in more than 160 countries, servicing, obviously, thousands of people across every continent and beyond. And I can imagine where sometimes same organization may stretch across those, but also your consumer side may experience that differently. And in every country, you’re going to have a different schooling system, you’re going to have different regulations, you’re going to have different states of play for where AI is. How are you managing that at Coursera, and does that provide more opportunities? Does it provide ways of trying to focus on some parts of the world in different—Vietnam in a different way than you might in the United States and somewhere?

Hart: Yeah, I think that our global scale and reach is a real asset for us. We have 300 million learners, registered learners worldwide. We work with hundreds of trusted institutions, thousands of enterprise customers worldwide. We believe that we have, among, if not the, richest data available about how the world learns. We also see how institutions are really focused on skill proficiency and skill development and how that differs across different institutions and different businesses. We support... Because we work with all those different parts of the ecosystem and support them in different ways, that gives us a really interesting vantage point that we can use in all of our conversations with existing partners across different segments and with new ones as well. And we combine that global distribution with very local content and insights in ways that we think are really powerful. One example would be SkillsFuture, which is a flagship national initiative in Singapore that is working in partnership with the Singaporean government. SkillsFuture is trying to really promote lifelong learning and to develop Singaporean citizens to their fullest potential. And so, Coursera courses are eligible for SkillsFuture credit. And so Singaporeans can access content from across Coursera to really help them build practical and very career-relevant capabilities. Another example is Kazakhstan, where we work with their ministry of education across all of their higher educational system. And so hundreds of thousands of students, through thousands of different, or hundreds of different universities are benefiting from Coursera. Another one, again on the government side, Mission Karmayogi in India. So India has the world’s largest, or one of, if not the world’s largest, civil servant organization across the country. And Mission Karmayogi is leveraging Coursera to help upskill and train public servants with the right skills so that they can then do a better job of serving citizens across India.

Kerr: Right. Well, thank you for taking us through so many parts of the organization. My final question is going to be much more just like a personal leader. This is September 2026. The uncertainty that surrounds AI—I mean, there’s a New York Times podcast a little earlier this week about the end of humanity in a few years, whereas others are seeing the jobs boom that’s happened with AI. So much uncertainty. What do you do, maybe personally, to kind of navigate through this uncertainty?

Hart: The diffusion of AI that we are very focused on through skill delivery, is, to me, going to be an amazing factor and phenomenon for humanity if we are able to manage the risks. And I think the risks are really important, so educating people on those risks is really critical. And then, of course, there’s a role for all of the frontier labs to play, as we’ve seen play out over the last week or so with all of the stories about potentially slowing down or pausing while we build the right safeguards. I think that’s always going to be a risk. And so, having the right human oversight of AI is really critical, because if AI doesn’t deliver value for humans, then that is a real problem. And so we are very focused on making sure that we help people gain the skills to get the most out of AI, but we believe that comes through the combination of humans and AI.

Kerr: It’s a great way for us to end. Greg, thank you so much for joining us today.

Hart: Thank you so much, Bill. My pleasure.

We hope you enjoyed the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you listen to podcasts. You can find out more about the Managing the Future of Work Project at our website at hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.