Meredith Whittaker on Who Controls Your Data in the Age of AI

16:36Meredith Whittaker:What do you think the risks are? If you're using Claude or ChatGPT, what do you think realistically the risks are over the next five or ten years that your data is compromised and some bad actor or the LLMs themselves will have access to your private information and be able to link identity with, I mean, the John Oliver segment on finding people's data. and the dark web, including their search history, should people be really, should people be cognizant of what they query these LLMs? I mean, I think they absolutely should be cognizant.

17:19Scott Galloway:Any query to an LLM that isn't sort of a specialized private inference setup, you know, kind of what Moxie, who founded Signal, is doing with Confer or other similar setups, but any, you know, a general query, chat GPT, is sending that data to servers that are controlled by OpenAI, Microsoft servers. They retain that data. They could leak that data. We know that when presented with a valid subpoena, they will turn that data over. In a world in which norms and laws and definitions of criminality shifts from one year to the next, perhaps. It's good to be cognizant of where that data could go and what it could do in terms of, you know, marking you as one or another type of person.

18:11Scott Galloway:Not to mention, I think, you know, with the introduction of advertising and, you know, increased targeting, at least the plan to introduce advertising in ChatGPT, I think there are also issues about what that can reveal about you, you know, in more mundane context as a consumer or as a job seeker and, you know, the kind of advantages or disadvantages that might accrue, given that the power to define you based on data that is, you know, in the context of ChatGPT, often extremely intimate.

18:46Meredith Whittaker:You've actually referenced that AI is a marketing term. What did you mean by that?

18:49Scott Galloway:Yeah, I mean, I think it's I'm being flatly literal, although I think that's sometimes taken to mean that I'm saying AI doesn't exist or it's not serious, which is, you know, marketing is in fact very serious. You know, what I'm talking about there is just sort of denaturalizing AI as a technical term of art. If you look back at the term AI, you know, it was created in, you know, 1956, 1957 by John McCarthy, who hosted the Dartmouth Conference. Those of us in this world will be familiar with that, a kind of an iconic conference where a number of the quote-unquote fathers of AI gathered to try to create intelligent life in the form of a machine over the course of a summer.

19:35Scott Galloway:And John McCarthy created the term in his own words in subsequent interviews because he wanted to exclude Norbert Wiener from the convening. They didn't get along. Norbert Wiener had created the term cybernetics in the field of cybernetics, and McCarthy classically did not want to be a disciple. He wanted to be the father of his own thing, a very common academic urge. And he also wanted grant money, and he thought artificial intelligence was a kind of flashy term with a cool valence that would get some of that Cold War era ARPA money flowing to his lab, which it did. It funded the conference. But over the history of the term, it's like over 80 years now, we've seen it applied to very disparate technical modalities.

20:21Scott Galloway:So McCarthy was invested in symbolic systems, which would look much more like decision trees and was actually deeply skeptical of the neural approach, which predated the term by about 10 years and was McCullough and Pitts and neural networks sort of stem from that. So what we see is a term that was invented primarily to describe an approach that's out of favor today has now been applied, you know, because of the specific resources available and the recognition that, you know, neural networks can do interesting things with data and compute and the type of business models we have. the term AI is now applied to an approach that was not actually under its umbrella when McCarthy invented it.

21:06Scott Galloway:And why is any of this important beyond it just being very interesting if you're a nerd? I think it's important because it allows us to step back and actually recognize that this is not a term of art. And what we are describing are very particular approaches that have their own historical and political economic formulations and that we can actually sort of have a bit more agency to define what we mean by intelligence, to choose the technologies that we are leveraging to produce intelligent-seeming outputs, and to be a bit more critical and actually regain a bit more of our own agency in relationship to mythologies that kind of naturalize these systems as just a sort of linear arc of technological and human progress.

21:52Meredith Whittaker:There's been a lot of, I don't know if it's warnings or catastrophizing from AI executives who said, I'm scared of what I've built and I need to retreat to the, you know, the Cotswolds and write poetry. I'm curious what you think the threat level is of AI and if it's been overstated, understated and where you see the biggest threats and how we as a populace respond to it.

22:18Scott Galloway:I think there are threats, particularly if we integrate these probabilistic generative and decision-making systems into high-stakes domains, nuclear, defense, energy, and put them to tasks that they are ultimately not secured or suited for. So, you know, you can have, you know, reward hacking, you can have emergent behavior. All of those things are real. Those aren't things that are simply going to sort of spring out of nowhere or, you know, Athena from Zeus's head and suddenly we have ephemeral technologies running around without our control or delegation in some sense. Those would need to be choices that are made by people and decision makers.

23:08Scott Galloway:And I do think, in some sense, some of the fear has a bit of escape velocity from material reality and almost sounds a bit like a religious fervor rather than kind of a technically grounded concern about the rush to integrate technologies that are not fit for purpose and could have collateral consequences, which is where I land on it. My primary fear, however, is the combination of the mythology of artificial intelligence, which is really framing these technologies as superior to human judgment, superior to human capabilities, which on some axis measured in some ways. sure, they do math much quicker.

23:51Scott Galloway:So does a calculator, they can produce things more efficiently, et cetera, et cetera. Yes, but ultimately, these are very centralized technologies that rely on huge amounts of data, data that is captured by an industry invested in what I'd call the surveillance business model, which is effectively collect all the data you can via your platforms and then train an AI model, sell it to advertisers, et cetera. And so it requires huge amounts of data. It requires huge amounts of infrastructure. And I don't have to go into the wild CapEx spending, the kind of NVIDIA's picks and shovels, the monopoly on chips and the build out of data centers.

24:38Scott Galloway:And it requires huge distribution networks, which often get left out of that calculus. But basically, if you're going to make money or you're going to integrate this, you need either a large social media or marketplace platform, or you need a cloud business model, or you need to latch on to one somehow. So all of that redounds to an industry that is highly concentrated in the hands of effectively the winners of the last tech boom, the platforms who were able to establish data pipelines and massive amounts of data, large platforms, cloud infrastructures, global reach that were sort of cemented via network effects and economies of scale, all classic communications network monopolies.

25:22Scott Galloway:And so my concern with all of that is that what we're looking at is a significant concentration of power over infrastructure and decision making that is then rebranded as a kind of God's head intelligence in ways that are making us less critical than we need to be about how that power is being leveraged.

25:42Meredith Whittaker:Well, let's drill down to specifics. What do you think, and nobody knows, but what is your best guess with respect to AI and employment, and let's call it the West, and Europe and the U.S. over the short and the medium term? I've seen TikToks of economists and AI executives saying, or AI thought leaders saying, employment, you know, we're going to see a massive destruction in the labor force. But the flip side is so far, it hasn't really manifested. There's some there's you could potentially interpret that the job market is softening, but youth unemployment is about where it has been historically at average.

26:26Meredith Whittaker:AI and the labor force. What is your best guess?

26:31Scott Galloway:Yeah. And I got to be careful here. This isn't really my lane. And I'm seeing a lot of kind of competing headlines. It does seem clear to me from some conversations that at least in part, AI has been a handy pretext for job cuts. You know, boards and media and shareholders will accept that, you know, hey, we cut X number of people because this is part of our AI strategy. That doesn't look like weakening demand. That looks like innovation. And so I do think there's some AI wrapping of downsizing that is happening. And I've heard that firsthand from some folks. I do think, you know, we are seeing at least a sort of degradation of work.

27:18Scott Galloway:And, you know, degradation meaning, you know, there are people who maybe used to have a job as a copywriter or a translator. And, you know, we've seen this with translation who are now just kind of editing AI output, right? And it's a less secure, maybe less fun, less rewarding job. But it's not removing the human. It's sort of removing the agency and power that a human would have in that job under different circumstances. I am really impressed with what I've seen. Or, you know, it's the new round of coding agents are very, very capable. And, you know, they're definitely seeing a lot of excitement across my industry there.

28:01Scott Galloway:You can't deny that these are very useful and produce output that is pretty commensurate with a junior programmer. But again, you still need a senior programmer. You still need somebody who understands how it works to review the code and maintain it. And so even though you're seeing advances in capabilities, one thing that isn't being talked about enough is there are a few things that many engineers I've worked with hate more than having to maintain it. someone else's shitty code. So you still need somebody who has an understanding of the systems level, who's bumped their head up against problems and understand them, you know, and can fix them, who understand how one, you know, pull request or kind of tranche of code might interact with another.

28:50Scott Galloway:And that's the place where I'm not only concerned that the kind of rapid outsourcing of some of the development work to agents. I think some of that could backfire in a kind of technical debt that is very difficult to pay down if what we're looking at is systems that are sort of built by agents or coding AI and not fully understood by the people who, the kind of skeleton crew who are left to maintain them. So those are some reflections. I don't think I have a clear answer because I think this is not just a question of AI. It's also, you know, where is their market will? You know, how is AI going to be used as a pretext?

29:36And then what happens when we do have the first significant issue with the reliance

29:43Scott Galloway:on these AI systems? And I, you know, I say that as I recognize that, you know, Amazon went down apparently because of an error made by an AI agent that they integrated. So, you know, we have already seen a kind of first wave of critical issues that are caused by a kind of dependence without human oversight.

30:08Meredith Whittaker:We'll be right back.

30:14Meredith Whittaker:Support for the show comes from VCX, the public ticker for private tech. For generations, American companies have moved the world forward to their ingenuity and determination. And for generations, everyday Americans could be part of that journey through perhaps the greatest innovation of all, the U.S. stock market. It didn't matter whether you were a factory worker in Detroit or a farmer in Omaha. Anyone could own a piece of the great American companies. But now that's changed. Today, our most innovative companies are staying private rather than going public. The result is that everyday Americans are excluded from investing and getting left further behind, while a select few reap all the benefits.

30:48Meredith Whittaker:Until now. Introducing VCX, the public ticker for private tech. VCX by Fundrise gives everyone the opportunity to invest in the next generation of innovation, including the companies leading the AI revolution, space exploration, defense tech, and more. Visit GetVCX.com for more info. That's GetVCX.com. Carefully consider the investment material before investing, including objectives, risk charges, and expenses. This and other information can be found in the Funds Perspectives at GetVCX.com. This is a paid sponsorship.

31:24Scott Galloway:This episode is brought to you by Nespresso. Introducing Virtuo Up, the latest in a long line of innovation from Nespresso. It's innovation you can touch, sense, and taste in every single cup. With a three-second start, easy open lever, and dedicated brew over ice button, it's even easier to enjoy your coffee your way. Sip for yourself.

31:46Meredith Whittaker:Shop Virtuo Up exclusively at Nespresso.com. I'm RJ Decker, a private investigator Uncovering the Sunshine State's Darkest Secrets Tuesdays, it's the premiere of ABC's Hottest New Crime Show RJ freaking Decker as I live and breathe He's a private eye It's not a standard murder, it's someone bigger And a public mess Trying to get sent back to prison today? You go to prison one time and suddenly it's All the jokes RJ Decker, series premiere Tuesdays on ABC and stream on Hulu.

32:23Meredith Whittaker:We're back with more from Meredith Whitaker. So there's a tension between privacy and encryption, and I think the potential weaponization of encryption and privacy by bad actors. And I would imagine by virtue of your position, I think I understand where you would land on this or at least a bias or a view on it. And in London and New York, they say you can't go more than 12 or 15 feet outside without being on camera somewhere. And to a certain extent, I like that. I think I like it more in Britain because I'm less worried about it being weaponized by the administration here. But if you look at the decline in crime rates, I think some of it is because of technology and then court-ordered, mandated, if you will, violations of privacy if there's enough evidence that this person is a bad actor.

33:15Meredith Whittaker:and that we need to violate people's privacy to understand, you know, if something bad is about to happen. You must be given this question all the time. That tension, where do you land on that tension? And is there ever a reason for why people's privacy should be violated in the context of larger safety concerns?

33:36Scott Galloway:I want to back this up to sort of the fundamentals of encryption. And when we're talking about, you know, signal, what we are talking about when we talk about into an encryption and the way that it works is a technology that either works for everyone or it works for no one. If you undermine the math of encryption, if you put a backdoor in there, you have a, you know, not actually random, random number generator, that means you could basically perturb the encryption, decrypt it. that's not just a backdoor. That's just not just an error that only the good guys can avail themselves of. That is effectively breaking encryption for everyone.

34:22Scott Galloway:So it really is a scenario where the people you hate the most have to be able to use it to exercise that right, so to speak, if the people you love the most are going to have access to it as well. It's all eggs in one basket. And that's at the level of math. I'm not answering the question, is it ever good or appropriate to undermine, you know, that's not actually what I'm talking about. What I'm talking about is a world in which over the last 30 years, we are surveilled within an inch of our lives. You said every 12 feet were recorded. Great. And then you made the comment, you know, I'm more comfortable with that under one regime than I am under another regime.

35:08Scott Galloway:Well, that becomes the issue. You're not really in control of, you know, how the sands of that regime shift. I mean, maybe you vote or, you know, whatever it is, but that data is indelible. Those systems are pervasive. Meta is adding facial recognition to their Ray-Ban glasses, right? Where is that data going, which governments will access that. And so it is interesting to me that in a golden age of surveillance, when unprecedented in human history, our actions, our preferences, our communities, who we date, who we talk to, what we do for a living, how we spend our money, are surveilled and logged at a level of detail unimaginable to the Stasi, that we are still pinpointing a tiny refuge where the fundamental right to private communication that is recognized as such, that is necessary for a full and joyful and intellectually rigorous life that has intimacy and the ability to exercise our opinions and dissent and blow the whistle and do journalism and all of that, that that one right is presented as a problematic and as the barrier between stopping crime and allowing it to run rampant in a world where, you know, the issue is more often than not, you know, finding the needle in the haystack of noise and the haystack of data, not getting access to an encrypted channel.

36:42Scott Galloway:So my stance on that is very, very clear, but I also think the framing of the problem needs to be shifted a little bit.

36:49Meredith Whittaker:Yeah, my Pivot co-host said something that really struck me. She said that people have the right to have secrets. And it really struck me. And the kind of the smartest people I know that also understand tech, I'll use Signal. And I realized how promiscuous and careless I've been with my own data. And I thought what I do is just not that interesting. And most recently, when I hear the Trump administration talking about assembling lists of people who are vocal, pretty outspoken against the Trump administration, I'm like, wow, I spoke too soon. And if you were to, and you have advised the government, I know you were part of, you worked with Lena Kahn.

37:28Meredith Whittaker:What regulation, if you were to advise the administration or the FTC, maybe it's under a different administration, on what would be the most thoughtful regulation as it relates to privacy, encryption, or AI, you know, kind of magic wand time. What do you think is most needed from our governments right now?

37:47Scott Galloway:Yeah, this is a bit of a tricky question for me because I've been not in the policy bubble for a while. I do think, you know, something as simple as a meaningful consent and by which I do not mean just a bunch of click wrap and cookie banners around whether or not a given company or institution gets to create data about us at all. Not what they do with our data, but whether they have the right to tell my story, to know about me, would go a long way. Of course, that would wreck an entire logic of the tech business model. But I do think the fundamental thing that needs to be done, however the regulatory paintbrush would paint this, is to question and then take back the authority to define who we are from a handful of companies that have naturalized their right to sort us and order us and tell us our place in the world.

38:56That's a bit of a philosophical answer, but I do think that's the core issue is the authority we've given tech companies who create data for advertisers to sort and order our world and tell our stories for us.

39:10Meredith Whittaker:I'm just curious what you thought of the Ring Super Bowl ad.

39:14Scott Galloway:My God. My God. I mean, I didn't expect it to become so flagrant so quickly, I guess. and seeing that it was, you know, I was like, who are they selling this to? Is it people who would install this or is it the government contracts who recognize exactly what this is selling and want to sign up for data access? Like I certainly wasn't the core demographic it was aimed at, but it also felt like there was a, you know, a tertiary market that was actually being addressed that wasn't, you know, eager doorbell owners.

39:52Meredith Whittaker:When you look at the landscape, well, I'm going to ask a market question and my guess is you're going to tell me it's not your lane, but I just want to remind you that's never stopped us from opining on it on all manner of topics. We have no domain expertise in. In the markets, there's been a meltdown around SaaS companies from a valuation standpoint. You work, you essentially work for or run a software company is the way I would think of it. I don't know if you call it that, But at the end of the day, I would imagine it's code.

40:21Scott Galloway:I work for a software company, yeah.

40:22Meredith Whittaker:There you go. So there's been an enormous destruction of value among SaaS companies believing that AI is going to come in and kick the crap out of these guys or make them obsolete. When you see that happening, do you have any initial thoughts on the viability of some of these software companies who are, you know, some of them lost 40, 60, 70 % of their value?

40:44Scott Galloway:I think ultimately when you're providing enterprise software, particularly to highly regulated industries, it needs to be interoperable with legacy equipment. Even if you don't like that legacy equipment, there's a superstructure there, there's a foundation. It needs to work with the data that you have, even if that data isn't great. That data needs to be cleaned and fungible. You need to be able to account for the different determinations that are made, depending on what kind of model you hook in there. That might not be a possible, particularly in financial services and other industries with high compliance burdens.

41:24Scott Galloway:You need to have often human oversight that is personally liable or accountable for different decisions. So I do anticipate that AI in some form or fashion will be integrated, will have impacts here. But fundamentally, this is not a magic wand, right? And there's a lot of legacy infrastructure, regulatory burdens, and labor processes and modes of work that need to be accounted for. And I don't see SAS software going away anytime soon. And I don't see AI doing anything to really erase those other considerations, right? I think predictions of the demise are a bit self-interested and far premature.

42:15Meredith Whittaker:And last question, Meredith. It strikes me that young people are absolutely, at least when I see their actual behavior, your, there's some consumer dissonance in that as people talk a big game about privacy, and then I see people basically telling the world where they are, what they are doing, and who they're doing it with. And it strikes me that even if you put a thin layer, if Uber would ever get hacked, a thin layer of AI on top could basically connote who's having affairs, terminating pregnancies, HIV status. It just wouldn't be that difficult to just know everything about someone with just their Uber data.

42:50Meredith Whittaker:Do you see the same dissonance I see in that as consumers have just decided to trade off massive privacy for utility? And do you have a message for them?

42:59Scott Galloway:I do see some of that. I would shift it a little bit. I think ultimately humans want to be loved and they want to be included. They don't, you know, even when we talk about signal and privacy, we're not talking about a vacuum. It's not Meredith by myself with none of my thoughts escaping the anechoic chamber of my meaning making. I am using Signal to share what I think with other people because I am a human and communication maps to human relationships and the desire to be connected and to be included, etc., etc. So I think we're in a world where ultimately we will opt as human beings. I use these services too because I want to go to the party.

43:46Scott Galloway:I want to see what people are doing. You know, I got to get somewhere. I want to participate in life while I'm living, as do I think most people, right? So the ways to do that are things we're going to do. And I don't think they represent actual choices about where we feel comfortable or uncomfortable with our data, whatever our data might be, right? We don't really have access to it. We know we don't want someone to share our mean DMs with our friend. We know we don't want our health data leased to our insurance in ways that would harm us. But that's also a place we don't have that much control.

44:27Scott Galloway:And in the meantime, we got to get to work. We want to see what our friend posted. We want to be part of the popular people. And the ways of doing that have been slowly, you know, we can say colonized or sort of, you know, instrumented by these tech services that advertise convenience, advertise connection, advertise ease, and then below the surface have sort of hollowed out our privacy and our ability to, you know, define ourselves in the place in the world. So I would say what we're seeing is a natural human inclination. You know, we use what we can to be together, to connect with each other, to participate in life.

45:06Scott Galloway:Those services have themselves, I think in some sense, betrayed us structurally. And that doesn't mean we don't care about privacy. That means a meaningful choice around what it would take to care about privacy has not really been given to us. You know, we do see the number of people using signal going up and up and up. We do see people's understanding of why privacy is so important. I think becoming more acute and more felt for people at a personal level when they see people's social media posts being used at the border, when they see these collateral consequences that are coming home. I think the issue then is, okay, what do we do about it?

45:49Scott Galloway:And you can't say, well, the choice is never to communicate with your friends because that's simply unrealistic and anti-human. But you should use Signal.

45:58Meredith Whittaker:I'm not exaggerating, and this is my final plug. The smartest people I know, and then the people who understand technology, the most have domain expertise around technology. That vent overlap, they all use signal. It's almost like a badge of like, I get it. You know, anyways, but my favorite quote from this is people are people want to be loved and included. Meredith Whitaker is the president of the Signal Foundation and a leading voice on AI policy. She co-founded the AI Now Institute at NYU, advised FTC chair Lena Kahn and was named one of Time's 100 most influential people in AI. She joins us from New York.

46:36Meredith Whittaker:Meredith, I very much appreciate your time and your good work. And I meant what I said. You're the bright, well-lit, clean part of the AI technology bookstore. Let's put Meredith Whitaker in charge. Let's just, let's consolidate all of it. I'll go raise$11 trillion, buy all of these companies and put you in charge. Deal?

46:57Scott Galloway:It's a deal, Scott. It's a deal. Yeah, looking forward to working with you. And thank you for having me on.

47:26Algebra of happiness, a hack for young dads.

47:30Meredith Whittaker:It is striking to me how selfish kids can be. I mean, it's just a I feel like I'm essentially a essentially a credit card that occasionally gets to watch a football match with them sometimes. And let me just give you a hack. If you're a dad like me who thinks that you're going to have all these hallmark moments with your child, you'll have some of those. But for the most part, it's going to be mostly a one way relationship. And I'm not saying it's not amazing, but the hack I have implemented and it's helped me a lot is that my favorite title, I've been a founder, you know, all these cool titles, whatever.

48:11Meredith Whittaker:My favorite title in the world is dad. dad and that is every time my kids call me or say oh hi dad or they call out dad or you know I love you dad every time I hear the word dad I'm like one of those dogs that hears the word walk and I've trained myself to just love that term it's the most important term in my life and it just it's more dope for me than anything is when these two things that kind of look smell and feel like me, called me dad. And what I've decided, and I started believing and training myself to believe five years ago, is that when my kids are awful, you know, they give me a hard time or they come home and expectorate their emotions or they're unreasonable or they slam their door.

49:01Meredith Whittaker:My kids, and what you'll find is generally speaking, your kids don't behave that way outside of the house. If you're like 90 % of us, you're going to find that outside of the house, your kids are pretty reasonable, pretty good citizens, pretty polite. And at home, they're fucking terrorists, assessing the household for vulnerability so they can strike when you're at your weakest. Now, why do they do that? Because they're processing, they're emoting, and they know what they can do with you because they know you are there unconditionally. They know you love them unconditionally. Why? Because you're their dad.

49:37Meredith Whittaker:And so what I have done, and it's been a real unlock for me, is that when my kids say something inconsiderate or even mean to me or aren't respectful or aren't kind, I'm not saying I let them roll right over me, but I assume they're saying one thing to me. They're saying doubt.

50:01Meredith Whittaker:This episode was produced by Jennifer Sanchez and Laura Genere. Cammie Reek is our social producer. Bianca Rosario Ramirez is our video editor. And Drew Burrows is our technical director. Thank you for listening to the PropG pod from PropG Media.

50:23Meredith Whittaker:Rinse knows that greatness takes time, but so does laundry.

50:27Scott Galloway:So Rinse will take your laundry and hand deliver it to your door, expertly cleaned. And you can take the time pursuing your passions. Time once spent sorting and waiting, folding and cueing, now spent challenging and innovating and pushing your way to greatness. So pick up the Irish flute or those calligraphy pens or that daunting Beef Wellington recipe card and leave the laundry to us. Rinse. It's time to be great.

50:52Meredith Whittaker:Thought Sweetgreen was just salads? Think again. There's a new way to do Sweetgreen. Wrapped and ready. These handheld wraps pack bold flavor and 40-plus grams of protein into something hearty, satisfying, and built for life on the go. From craveable sauces to satisfying textures, they're designed to keep you going without slowing you down. So put that fork down. Try the new wraps today in app or at order.sweetgreen.com. Available at participating locations only. 30 years ago, Blinds.com broke the mold and made custom window treatments easy for everyone. Over 25 million windows later, we're celebrating by giving our customers up to 50 % off site-wide during our anniversary sale.

51:36Meredith Whittaker:Whether you DIY it or want a pro to handle everything from measure to install, Blinds.com has you covered. Shop online, access real design professionals, and get free samples. Thank you for 30 amazing years. Shop the anniversary sale now through March 11th and get up to 50 % off site-wide at blinds.com.