Reid Hoffman on Building AI and Other Tech More Responsibly

Legal teams face more data and more scrutiny than ever. They need AI built for both. Relativity is the AI platform for legal work, delivering defensible AI that handles the tedious tasks so judgment stays where it belongs with you. Learn more at relativity.com/ideacast.

Marketers know how it feels to invest in something that seemed incredible but didn't live up to the hype. We optimize for impressions, but then don't see any revenue. Instead, get the highest ROAS of all major ad networks with LinkedIn Ads. Spend $250 and get a $250 credit. Just go to linkedin.com/ideacast. Terms and conditions apply.

On May 20th, join me at HBR's Annual Leadership Summit with masterclasses, interviews with the CEOs of AT&T and Mattel, and an interactive case discussion led by Harvard Business School Professor Karim Lakhani. For 25% off, be sure to grab your early bird ticket by April 14. To learn more, go to hbr.org/leadershipsummit. See you there.

Welcome to the HBR IdeaCast from Harvard Business Review. I'm Alison Beard. When any new technology comes along, people, especially those in the tech industry itself, tend to get really excited about all the good it's going to bring. Social media connects people around the world, crypto democratizes finance, generative AI supercharges productivity, and so on. The evangelist crowd is loud and proud. But as we've seen over the past decade, the potential downsides of the latest tech innovations don't always get as much attention. Yes, you'll see some skeptics warning about unintended consequences and negative externalities, but it doesn't seem like industry insiders, the people building and deploying these new tools and the leaders overseeing that work, are thinking all that hard about what challenges they might inadvertently create. Our guest today is an unabashed techno-optimist. He really does believe in the power of technology to improve our lives, but he also knows how important it is for tech companies to think more carefully and responsibly about the problems they're trying to solve and the products and services they're putting out into the world. Reid Hoffman is a founding board member of PayPal, a founder of LinkedIn, a partner at the venture capital firm Greylock, and a director of several companies, including Microsoft, although he recently stepped down as a board member of OpenAI. He's also a podcaster hosting Masters of Scale and the new show Possible.

Reid, welcome.

Great to be here.

Okay, first off, how do you define responsible or ethical technology?

So one of the illusions that is sometimes promulgated is that technology is essentially value-neutral. And that doesn't mean that it embodies values in kind of a simple way that, like, you know, I believe in democracy, or I believe in some other form of human organization, you know, or kind of the various values debates we're having within the US and other countries. I think that the question is, you say, well, how does this affect the human condition? What does it mean for different individuals? Like, are there bias issues? Are there things where it creates some kind of bad social impact? And you have to ask these questions. Obviously, one of the challenges when you're dealing with things of scale is it's never all good, like 100% everything. I think what you have to do is you have to make it on broad, really good, and then try to make sure that you're not disadvantaging groups that don't have, you know, power or a voice. So for example, you say, well, cars— well, cars are generally speaking very good. It enables transportation, enables mobility, enables people to live in different areas. On the other hand, of course, in the US we have 40,000 deaths per year in driving, then of course climate and all the rest. So you have some kind of challenges, and you try to shape it so that on balance is very good and you're dynamically improving as you learn and refine.

As someone who has been a leader in the tech industry for a really long time, what is your honest assessment of the job that you all are doing in considering not just the upsides, but also the downsides, and then trying to mitigate those risks, whether, you know, that's social media a decade ago or generative AI today?

Well, you know, it's a little bit hard to talk about the entire tech industry because there's some people I think who are doing pretty good jobs, and I think there's some people who are doing pretty awful jobs. You know, the story of social media, as you said in the intro, is when it opens with blogs and social networks and all the rest, it's like, oh, we're giving voice to the people who didn't otherwise have voices, and people who might be a minority of some sort somewhere in the world, whether it might be, you know, kind of sexual orientation, or might be religious, or might be a racial minority. They can discover their voice and they can connect with other people. Isn't that awesome? And of course it is, and it continues to be. But then you say, well, now it becomes where everyone's there, and then all of the issues that become part of why we have government, why we have regulation, and how we make society work together, those then come in place in full. And like, for example, one of the classic things that I've been debating since as long as I've been on basically television— I think I did a 1996 firing line on this— on freedom of speech is to say, well, we don't regulate freedom of speech. And it's like, well, of course we do. We have truth in advertising. We have issues around hate speech or violence, or there's all kinds of ways we regulate speech. Many of you say, well, my freedom of speech allows me to say false advertising and to sell drugs that are harmful for lots of money, you're like, well, that we don't allow in society. And I think that's what the tech industry is still coming up to speed on in terms of how do we navigate what is our definition of truth in collective discussion and how do we navigate that. Now, when it gets to AI, which is obviously the thing I've been spending a ton of time on in the last number of years, I think the tech industry has learned from the social media side to pay more attention here. So the questions around, well, is it biased or might there be unintended consequences in jobs or misinformation? The way that ethics starts is by asking the questions and checking as you're building. You're not gonna get it perfect. You're not gonna launch something at scale and get it perfect. But if you're asking the questions and you're measuring and you're improving, then you'll eventually get to a very good place.

So it seems like you're saying that technologists today and leaders of tech companies have maybe learned from the era of social media and that the famous move fast and break thing era is over?

What I'd say is, okay, you know, I'm the author of Blitzscaling. I'm a definite move fast person. The question is, what things do you break? You know, you break your servers, fine, no problem. You break society, no, that's a problem. And what I'd say is some tech leaders, you know, Satya Nadella, Sam Altman, are well on the learning curve. I think it'd be a fool's statement to say, oh, we've learned, we're good. It's like, no, no, part of what you're doing is we're exploring this new stuff and we're building these new things and you can't pre-predict all of it. You're learning as you go and you're fixing.

What factors are they considering when making business decisions? You know, what does the general public maybe not see or hear about what's going on behind the scenes, both in the VC community and within the companies themselves?

So I'd say that every technology company that I'm a part of, that my partners at Greylock are part of, are all at least asking the questions and doing it as part of how they develop the technology. And the questions can range from, are we being responsible stewards of data and people's trust? Might there be groups that are being disadvantaged by this technology that is a structural bad thing, like for example, you know, racial disadvantage? Whereas you say, well, we're We're disadvantaging criminals. Okay, that's fine. You know, fraudsters. You know, as best we can, are we red teaming and thinking about blind spots or things that could go wrong? Are we thinking about what happens when this gets to scale? Do we have a good theory about why this will be net really positive and how we can remediate or diminish harms? I think all of those questions in every tech company that I'm part of are central. And, and we go out and learn. We hire people and, and ask for, you know, what are the other things we should be thinking of doing here?

Yeah. And bigger picture, is the industry— and I'm sorry for keep saying the industry as a whole, but most of the people you know, you know, are you now focusing on more important problems than perhaps you once did? You know, there's the famous Peter Thiel quote, we wanted flying cars, they gave us 140 characters. And now it's more like we want climate change solutions, but we're getting a chatbot that can write stanzas like Shakespeare.

Well, yes, and, and complicated. So for example, flying cars. I'm on the board of Joby. We are working on flying cars, and that's to redefine space in a climate change way that will help with gridlock and pollution and a bunch of other things and be accessible. On the other hand, The natural pattern of these things is to try to figure out what's the easiest work to do that's most valuable. And so that's why people tend to do a lot of software. And I tend to think that actually chatbots can be really valuable. They can be valuable for anything that ranges from, "Get me some good information," to, "Help me solve this problem," to any number of things that could play into human life. But on the other hand, of course, solving hard problems like climate change, ocean deacidification, you know, other kinds of things are super important and people are working on those. They're just harder because it's a lot more expensive with the economic rewards being much more challenging. One of the things that, you know, I try to give thinking to and advice to is how do we create an incentive system that also goes after the hard problems more.

Yeah, absolutely.

And you use the word valuable, so let me press on that a little bit. By valuable, do you mean valuable to society, valuable to investors? You know, where is that purpose-profits trade-off or balance falling for your community now?

Well, in an ideal system, you align them so that the high functioning of business where the product it's offering their customers, it's really good for the well-being of the customers and society and the stakeholders that they're in. There are, of course, places where that gets misaligned. And it's not only within the tech industry, it's, you know, this is one of the challenges we kind of have with making industries work. And look, there is a— in all of society, there's a whole bunch of people who are doing things only for money or only for profits. That's part of our how we design the alignment of society that goes all the way back to Adam Smith. But the question is, is also that people will say, I want to hold my head up with my friends and my community and say, I'm doing a really good thing. All the people I hang out with are focused on how is it that we're also making the world and society better with what we're doing. And so for example, that's one of the questions we ask at Greylock when we're investing is to make sure that we are positive on those vectors. Now, we have to do so within the context of a strong business. But, you know, if you're asking the question and intentionally trying to do that, then that's at least half the game.

And so as a VC at Greylock, you know, what are you looking for, seeking out both in business ideas, business models, and founders right now?

Part of the thing that's a delight about venture investing is while you may have a very active theory of the game— so I've been doing, you know, generative AI for the last few years. I co-founded a company called Inflection with Mustafa Suleiman. We have Adept and Cresta and Snorkel and all these other companies at Greylock. And so we have a very active thesis on artificial intelligence and have had for, you know, 5+ years. We're also being surprised by the amazing things brought to us. So, you know, just to kind of illustrate what I think the quality of, of being surprised is, is when Brian Chesky and Nate and Joe brought Airbnb to me, I hadn't really been thinking about, you know, a marketplace for space, a question about how you can not just travel to a place to see a monument, but to experience local culture, to enable people to transform their own economic outcomes of being able to afford their house or their space. But yet that's just software, and that brings all that together. So, you know, for me, in addition to AI, I also tend to look at networks that redefine our social society space. It's part of the reason I created, you know, LinkedIn with my co-founders, things that we've done in various other investments at Greylock, including like, for example, you know, like take Roblox, which is, you know, okay, you've got developers building entertainment and educational things that generally speaking, mostly appeal to kids, but a whole range of experiences. We're looking to be surprised. And the question we ask is, are the customers net really benefited from this? And are the, is the community and society that they're in broadly also benefited? And does it have a very strong business that will transform industries? And, you know, if we see all that, and we see an entrepreneur that we think is high integrity, and that we would be delighted to be in business with our entire lives, then we get really excited and join forces.

Yeah, that high integrity piece, you know, finding people, founding teams who are absolutely trying to scale and run with their ideas and make a change, but then also will take that moment to step back and, and ask the questions about, you know, ethical construction, deployment, etc. How do you evaluate for that?

Well, it's not a simple formula, but one of the things we do pretty rigorously is reference checking. You haven't completed your reference checks until you found a negative reference check on everybody in the world. So for example, if someone was reference checking me in depth, what they would find is, oh my gosh, he's a really great creative problem solver, but he's not particularly good at making the trains run on time, right? And obviously, you know, when you're asking the integrity question, you're asking a question of how much do you actually in fact walk the walk, not just talk the talk? How much do you, when you get in positions of stress, do you make decisions, like for example, that say no, no, yeah, that would be the easy decision, but that takes risks and other people's well-being. Let's take the hard decision. Do you honor your commitments? And therefore, you know, when you're saying, hey, we're going to have a commitment to make sure that we are tracking how we impact society, we're going to have dashboards on it, and we're going to be improving them. Year by year. Will you be doing that?

Yeah.

Talk a little bit about the role that the tech world, the VC world, an industry that is still very much dominated by rich white men, has to play in increasing inclusivity and also decreasing socioeconomic equality.

One of the things that I've been saying for maybe a decade plus now, anytime you look at a problem, you go, that's important to solve, you go, if you're not part of the solution, then you're part of the problem. So you need to be saying, how am I as an individual and also, of course, as a firm and everything else, investing in trying to solve this problem? How am I putting in sweat and blood into trying to make this happen? And so relative to diversity and inclusion, in making sure that you have a regular workflow and process by which you're trying to recruit, you're trying to meet entrepreneurs. I mean, we do things at Greylock like have a set of office hours that's only for underrepresented minority entrepreneurs. We do in any recruiting thing, we make sure that we are interviewing disproportionately large numbers of underrepresented minorities, including unfortunately venture women, which is like, well, you know, aren't they half the population? Like, yes. And doing everything you can. And so, for example, we've helped stand up kind of new venture firms because when they come to us and say, hey, we're going to— we think that one of the things may just be having a venture firm that's entirely focused on funding women entrepreneurs might be a good way of doing it. Great. We'll help you. Right? And so you have to do all that kind of stuff. And would I want the progress to be 10x faster than it is going? Absolutely. And if someone figures out a way to make that happen, we'll help, we'll support. On the economic gaps, it's always a little tricky 'cause it's dynamic over time. For example, one of the things that I do, it was the same thing in my philanthropy as I do in my investing, which is you find an amazing entrepreneur, In this case, it's Byron Auguste, who says, look, there's all of this massive growth in the tech jobs and tech industry, and we wanna make sure that it works for the communities of color, works for women, works for, you know, other minority groups. Let's go make sure that a whole bunch of these people have pathways into tech jobs and make that happen. And so they at least can begin to bring their families in, understand kind of what the tech opportunities are, have, their communities begin to be able to benefit from participating in these industries. But by the way, when you're growing a new company, the new company makes the executives and the founders the most money, and it makes the next group of people the next most money, and et cetera, et cetera, is the way it works. So it doesn't necessarily immediately cause distribution economics, but you're trying to get everyone participating, and then you're trying to make sure that that the next generation of founders has the diversity that we have in society.

Ever invest in something that seemed incredible at first but didn't live up to the hype? Marketers know that feeling, optimizing for impressions and then not seeing revenue. Instead, invest in what looks good to your CFO. LinkedIn Ads generates the highest ROAS of all major ad networks. Spend $250 and get a $250 credit. Just go to linkedin.com/ideacast. Terms and conditions apply.

You mentioned other industry leaders that you respect and admire, who you think are modeling good leadership, you know, not only on the, I'm running a great business, but also on the I'm working to improve society front. But, you know, the poster boys for the tech industry— Elon Musk, Jeff Bezos, Mark Zuckerberg— you know, they definitely aren't perceived that way, no matter how much money they might give to charity or how many rockets they might launch into space. Do you get the sense that the good guys, as Kara Swisher might call you, are developing as many acolytes as the people who sort of do still cling to that move fast and break things ethos?

Well, you know, and I myself argued with Mark Zuckerberg about, you know, kind of freedom of speech issues and other things. But like, for example, one of the things I do with him is the CZI Biohub, where he's trying to cure infectious disease for people all around the world and putting in a lot of money to that. And that's You know, because he's such the poster boy for other, you know, intelligent criticisms, he doesn't get as much credit, neither here nor there, for all this other amazing stuff he does. And I just kind of feel it's important to make that gesture.

Yeah. And there's no question that, you know, a lot of the people who make a lot of money then do a lot of good. You know, I guess it's just trying to marry the two is what we're talking about.

Yeah. Well, that's important to do too. But like, for example, there is a differentiation between people who go, all of my economics is for my own self-glorification, and people go, look, I'm making a bunch of economics and I'm also doing a bunch of things that I'm caring for a bunch of communities that has nothing to do with my self-glorification.

Yeah.

And, you know, I just— I, I say that in part because it's too easy to get on the criticism bandwagon, and I You know, I think it's important to note. Now, I'd say that the folks who are perhaps not beating these drums as extremely tend to have less— I think the word you used was accolades. You know, I think it's because it's the principal way that you get accolades is by defining something pretty extreme and beating that drum. And then people who, you know, think that you're the messiah for beating the drum in that direction. Then come follow you. If you're kind of like measured and saying, you know, like things I've been saying here, which is, look, it's a net benefit is the goal. I think you do have to move fast. I think you have to build things quickly. I think you will break things, including things that you don't want to break in doing it. I think you have to do it with care and attention. But I think if you don't do it with speed, than the people who do it with speed, who don't care about what the impact is, you know, set the rules. And so I tend to think that, you know, it's less good, call it media coverage, to talk about the people who are trying to be thoughtful than the people who are being extreme.

Do you think that Silicon Valley still sort of leads the world in terms of what the tech industry is thinking about, or do you see sort of different ecosystems developing their own ethos around purpose and profits?

Well, I'd say the two areas in the world that are the most tech-leading are both Silicon Valley and a set of cities in China, mostly along the coast. I try as much as I possibly can to help create other tech innovation centers in other areas of the world. Like, I was just in Italy, France, and the UK, you know, high-principled democracies that have a really good concept of kind of like what the human rights should be and so forth. I try to help as much as possible in facilitating the creation of entrepreneurial bases and tech industries, but I do think Silicon Valley continues to be along with the, like, we learn a whole bunch of stuff from China, the kind of driving drumbeat. And it's one of the reasons why I think it's a very good thing that the discourse is, like, I'm at dinner parties in Silicon Valley where, you know, part of the discussion is to say, well, how do we, now that tech is continuing to have larger and larger impact, what is the way that we make sure that we're doing the right thing?

Let's talk about China. Are those questions being asked over there also?

Well, not being a native Chinese speaker and not having been there for a few years, um, you know, I would say I think any group of people— hey, you've got a million people, you've got a distributed smart people, you've got a distribution of ethical people, you've got a, you know, a whole bunch of different things— I would say that their environment is more tuned at the moment to the, as it were, the rise of China and the success of the business, and somewhat less to, you know, like, for example, you know, what does this mean for disadvantaged minorities within society? You know, in China, I don't think you have any discussion in the tech companies of like what it means for the Uyghurs, what it means for, you know, other kinds of things. I think people are people. I'm not saying anything about the quality of the people in doing that. I just think it's the environment that they're operating in.

Are there opportunities for more collaboration, interaction, knowledge sharing?

So, for example, one of the things I've been, you know, highly focused on, along with the OpenAI and Microsoft folks, which is like AI safety and making sure that when you build these new, very large, very capable systems, that the net impact is very good, that there are no really bad impacts. And you say, okay, well, how do we make sure that the work that we're doing, even though, you know, we've put in a whole bunch of work and energy and cost and, and hiring— I think there's hundreds of people at Microsoft who work on AI safety— how do we essentially just distribute it for free? How do we offer it to everybody, including our competitors and so forth in China, in order to try to get to good places? Because that's part of being intentional and good people.

So I do want to turn to your new show. It's a very interesting, you know, your sort of addition of the show Possible to your Masters of Scale franchise, because one is sort of the founders, the entrepreneurs, the leaders of companies who, you know, made it big basically. And then Possible seems to feature people behind the scenes working on these really difficult problems you mentioned earlier, you know, so on democratizing higher education through technology, nuclear fusion, you know, to help solve some of our climate issues. Talk about why you wanted to launch the show and focus on those people as opposed to the famous corporate leaders.

So, you know, one of the things that I see a lot in the U.S. and see in some places of the rest of the world is what is referred to as techlash, which is more negativity and uncertainty about what technology is bringing versus the positive sides. And I believe, as a hypothesis but very strongly, and can argue for the— whatever scale problem you're trying to solve, whether it's climate change, whether it's economic justice, whether it's criminal justice, other things, 30 to 80% of the solution is technology. What I mean by that is Technology changes the scope of what's possible. It changes cost curves. It changes what you might be able to pull off with the resources that we have. We can help solve these really fundamental problems with technology. And it isn't technology is the only solution— part of the solution. It's also how we organize ourselves as a society, what we value, what we invest in versus other things, other kinds of things. But technology is an essential part of making that scale solution work. And so we want to go to essentially the leaders, the innovators, the imaginers of what the world could be in this really good new way, and to talk to them and to share that sense of, Here is where we should row towards. And I think we can, for example, solve these really big problems, climate change. And oh my gosh, we could build a world that's so much better than we are today. Let's get to it.

Does the new generation of founders seem excited about that, even if it means, you know, their big paydays might be 2 decades in the future as opposed to, you know, becoming a unicorn within 5 years?

Well, again, I think some are and more, more are, right? It won't be all are. It won't be, you know, some people will still be, you know, creating, you know, I'm not going to— I try not to throw entrepreneurs under the bus, but, you know, various things that I go, well, that's not a particularly great thing to create.

Delivering liquor to your front door or something along those lines.

Whatever, whatever the thing might be. You know, I guess the one I most often pick on is Juul, but like creating electric cigarettes or vape things, I think are net not positive. But go and have the imagination that through entrepreneurship, through technology, through invention, you could solve these things. And as we get— there's a super amount of very talented people in the world and we just want more of them working on these problems and thinking about the fact they could make a difference. By creating a technology, a business, a project that could focus on this and make it work. And that's the dialogue we're hoping to increase in the, you know, kind of applying our imagination to how we create the future.

Yeah, we haven't yet talked about the role of government in innovation, you know, and in regulation. So some of the greatest technologies, GPS for one, stemmed from government investment initially. So do we need more of that is part A of this question. And part B is, where do you stand on regulation for emerging technologies like generative AI? Should there have been more regulation on social media, et cetera?

One of the things when people say, For example, what do I believe that most Silicon Valley or a lot of Silicon Valley people don't believe? It's like, actually, in fact, government's absolutely essential. It helps create a lot of things, helps create not just rule of law and a society and healthy functioning economy, but also baseline investment in universities and technologies. And, and so I'm a big believer in those. I also think that regulation can be an important part of that. One of the challenges of regulation is that the baseline conception of how most people tend to think of regulation is to ask for permission, not forgiveness, to tell you that you continue to do things the way you've done them in the past, and you're kind of locked in with very slow change from that and tend to be done by people who don't necessarily understand what the innovation clock looks like. And so the principle that I usually articulate here is when is bad regulation better than no regulation? And by the way, the answer is not— that's not a rhetorical question— or never. Because, like, for example, when you get to the financial system, the absolute necessity of the financial system continuing to run, you say, well, actually, in fact, bad regulation is better than no regulation to make sure that the banking system doesn't break. And other kinds of things, that's because it's just too critical otherwise. Now, when you get to a lot of technology and you say, well, you're enshrining the past, the problem is, is, well, if the actual solution is technology in the future, then a regulation that that particularly slows you down or anchors you to the past will be potentially more damaging to humanity than not doing it in various ways. And so, well, do you say no regulation? I was like, of course not. But what you do is start by defining like, what are the outcomes that you're looking for? And can you set those outcomes to essentially the innovators, the companies, the other things, to say more of these outcomes and less of these outcomes? And we'd like to see a dashboard, we'd like to see it. Tracked by your auditors. And so, for example, when we'd like to see less violence on video, so do you say, well, I'm going to have a regulation to say you must have a 5-minute delay between uploading the video and the broadcast of it? And you say, okay, well, that may not actually solve your violence and video problem because, you know, terrorists or whatever else might trick or hack the system for it, and your regulation really just created a whole bunch of processes that didn't do anything. Whereas what you said to companies is said, well, okay, I recognize you can't get to zero because again, large-scale systems. But let's say for the first 100 views, it's a $1,000 fine. For the next 1,000 views, it's a $10,000 fine. And for every view after that, it's a $100,000 fine. You figure out how not to show, you know, murders on video. Right. And that's what I mean by defining outcomes in ways, and then having the innovation do that. And that's the kind of thing that I think is the pattern that we need to apply when it gets to technology.

So what advice do you give to people who are early in their tech careers right now? What are some of the pitfalls to watch out for, and how can they become great, more responsible builders of technology?

I think everybody needs to think about their own life path with a toolset of an entrepreneur. Doesn't mean they have to be an entrepreneur. I think another thing is to realize that the, you know, all the way back to the beginning of our discussion, that the creation of technology can be itself a great good if you're asking the right questions. I think that even, you know, questions when you say, well, you know, obviously like take an area that's fraught with a whole bunch of things, genetic modification, genetic engineering. So, oh, that could be really bad, obviously. But of course it could be really good, getting rid of genetic diseases in ways that just cause suffering. So if you're asking the right questions and you're doing it the right way and you're thinking about how do you shape it the right way, you can have a scale impact in the world that leaves humanity much better because of your effort. And I think that's the, you know, ask the right questions and help create the future.

Reid, thank you so much for being on the show.

My pleasure. Thank you.

That's Reid Hoffman, entrepreneur, investor, and podcaster with the new show Possible. And we have more episodes and more podcasts to help you manage your team, your organization, and your career, including an upcoming IdeaCast bonus series about how artificial intelligence will change work. Find them at hbr.org/podcasts or search HBR in Apple Podcasts, on Spotify or wherever you listen. This episode was produced by Mary Dooe. We get technical help from Rob Eckhardt. Our audio product manager is Ian Fox, and Hannah Bates is our audio production assistant. Thanks for listening to the HBR IdeaCast. We'll be back with a new episode on Tuesday. I'm Alison Beard.