AI Has Made Everything Worse | Hany Farid on the Internet's Truth Problem

Hany Farid, one of the world’s leading experts on deepfakes and digital forensics, joins us to explain why AI-generated content is pushing us toward a world where we can no longer trust what we see, hear, or even who we’re talking to online. From AI-generated misinformation and social media manipulation to voice cloning, deepfake video calls, and North Korean IT workers infiltrating major companies, Hany breaks down how rapidly the threat landscape is evolving. He explains why humans are not good at detecting deepfakes, how digital forensics experts actually identify manipulated media, and why organizations need to rethink cybersecurity, remote hiring, and digital identity. This episode also dives into Big Tech’s role in the crisis, whether regulation and liability can make technology safer, the growing influence of large language models on how we access information, and what it will take to rebuild trust online. Despite the risks, Hany remains optimistic about AI’s potential. The challenge, he argues, is learning from the mistakes of the social media era and finding a way to harness AI’s power without sacrificing safety, security, and trust. Subscribe and like this video for more conversations with the world's leading thinkers on AI, technology, innovation, and digital transformation. In this video: 00:00 North Korean IT workers are already inside companies 00:41 Welcome to a “trust no one” digital world 01:49 How much of the internet is already fake? 04:59 Are we heading toward the dead internet? 07:06 What happens when we can’t agree on facts? 08:20 The existential threat of a post-truth world 09:49 Could AI control the information we see? 11:33 Who is manipulating us online? 13:20 How AI is supercharging scams and fraud 15:41 Why you can’t reliably detect a Deepfake 17:28 Should Big Tech be held liable? 19:37 Hany’s Advice: Get social media off your phone 21:26 Can liability force Big Tech to change? 24:25 Should kids be banned from social media? 25:37 The real problem with social media’s business model 28:08 Will Big Tech control the future of AI? 31:20 “China will beat us” and the case against AI regulation 33:31 How experts actually detect deepfakes 36:16 How real-time face swap deepfakes work 39:39 The Deepfake threat facing businesses 40:25 AI fraud is already costing companies millions 41:47 How fake workers get hired 42:49 Why voice biometrics are no longer secure #ai #futureofwork #technologypodcast #cybersecurity #deepfakes #digitaltrust #bigtech #aivoice #digitaltransformation #futureofjobs #futureofai #tedtalk Connect with Hany: LinkedIn: https://www.linkedin.com/in/hany-farid-40a97935/ Our links: Visit our website: https://www.infotech.com/?utm_source=youtube&utm_medium=social&utm_campaign=podcast Follow us on YouTube: https://www.youtube.com/@InfoTechRG

• 0:02Speaker 00

  We know that the North Koreans have penetrated Fortune 500. We talked to one company who I won't mention by name. It's one of the largest employers in the world. They caught 1,800 potential North Koreans. That's not a typo, 1,800. And that's the ones they caught. Now you gotta ask yourselves, how many got through? So these are not hypothetical threats. We talked to one US defense contractor who had five North Koreans in their organization. Five. That's insane. So I don't think there's a single Fortune 5 company that probably hasn't gotten hit.

• 0:36Speaker 01

  This is a show about the future of tech and the future of work. I'm Jeff Nielsen, and today we're diving into deep fakes and living in a trust no one digital world. My guest today is Hany Farid, the world's leading expert on deep fakes. He's a total tech rock star and always amazing to listen to. The former dean of UC Berkeley's School of Information. He's now co-founder of Get Real Security and a forensic consultant to everyone from the New York Times to DARPA for how to uncover online fraud. I want to ask him what he's seeing out there. How are we being manipulated? What's really at stake? And what can we do to protect ourselves when we can no longer believe our eyes? Let's find out.

• 1:22Speaker 01

  Hani, thanks so much for joining today. Really happy to have you here. So let's jump into it. I want to talk about the era of deep fakes or, you know, it kind of feels like we're entering into, you know, what I've been calling a trust no one era. And I imagine you're thinking something similar, but just to sort of set the table, you know, if you're online, if you're using social media, what's your sense in terms of what proportion of all media these days is AI generated or faked in some way?

• 1:49Speaker 00

  Yeah, I mean, look, let's be honest. Before the world of generative AI, social media was not a beacon of hope and trust in truth. So let's, let's not pretend that we're fundamentally talking about an AI issue. Social media has been a problem for a long time because the business model of social media is user engagement, right? That's the business model. It's not truce. It's not honesty. It's not decency. It's how do I engage users for as long as possible? And it turns out we're sort of a bunch of jerks. And the absolute lowest common denominator of human nature engages us. And we all know this. So now I'm going to get back to your question. First of all, answering these questions is very difficult. It's even difficult for the platforms. Meta doesn't even know what's happening on their platforms.

• 2:34Speaker 00

  TikTok doesn't know what's happening. But what I can tell you is it's not so much what percentage of the content is fake, AI, untrustworthy. It's what percentage of the content that is viewed. So I understand there's a long tail on content that's uploaded. Lots of it gets one or two views. I don't care about that stuff. But if you look at the stuff that drives engagement, the millions, tens of millions, you're probably seeing upwards of half is now fake. Either fully AI generated, it's an AI influencer, it's somebody co-opting another influencer. It's state-sponsored disinformation campaigns from Russia, Iran, North Korea, China. It's people trolling, it's ideological, or it's people simply trying to monetize.

• 3:19Speaker 00

  And the reality is, is that if you are using social media as your prominent source of information, and I don't mean going to the BBC's, you know, Facebook or TikTok or YouTube page, I mean, literally using social media, you are being asked to be lied to. You are, and we should be honest about that. It is not a place to get reliable information. And here's the thing, it never was. It's not what it was designed for, and it's not good at it. So I would say AI has made everything worse because it's fully democratized access to very sophisticated technology that used to be in the hands of the few. And now people can create fake accounts, fake influencers, fake content, fake images, fake video, fake audio. Everything is fake. The text is fake.

• 4:04Speaker 00

  The whole thing could be running with an agent. I can today on my laptop, run, run an agent that is just on social media all day long, 24/7, never gets tired and drive engagement. And then here's the other thing, last thing I'm going to say about this is these, the AI, what's so great about the AI or what's so damaging about the AI is it learns. It's like, oh, I did this and it didn't engage. Okay, I did this and it didn't gauge. So think the sort of A B testing that everybody does. The AI can now do this in a very rapid loop. So I would say probably more than half. It's probably worse than that. And here's the thing, it's only going to get worse. All right, the platforms are struggling to get a handle on this. YouTube did announce that they were eliminating over 100,000 channels that were just AI slop, as we call it.

• 4:49Speaker 00

  So I think that the, the platforms are trying to get a handle on it, but the scale at which they operate and the complexity and sophistication is going to make it very hard for them.

• 4:59Speaker 01

  I mean, to start, I guess lots of reactions to that. The, the number you threw out there of it being a coin toss is, is really shocking. And it's shocking, not just because it's so high, but because when you multiply that by how much content people are exposed to, like it, it really is a staggering number. When you think about, you know, just drowning in this content. And I mean, one of the things I worry about sort of philosophically, I guess, is, you know, you've probably heard somewhere along the way that the idea of like dead internet theory. But the idea that like, if I'm using Instagram or I'm on Reddit or a social media platform, how far am I from a world where if I'm not interacting with my friends, it could be all fake.

• 5:41Speaker 01

  It could just be some sort of Truman Show where it's just, this is all just farming me, plugging me into a machine to keep me engaged. It feels terrifying and.

• 5:52Speaker 00

  Turns out the Matrix, the Matrix wasn't science fiction. It was a cautionary tale because, I mean, what you're describing is they sent you the Matrix, right? And if, look, you know, do I think we're at the Matrix? No, but walk around and look at how people are on their devices scrolling like, like, like little monkeys. I mean, we're not that far off. And the, the paradox, of course, of the internet was it was meant to democratize access to information. And it did, but it didn't distinguish between good information and bad information. And so I think you're right. I do think that if you are spending your time, the majority of your time on Reddit, on TikTok, on social media, you are living in an alternate reality that is created for you.

• 6:37Speaker 00

  So this is the thing you have to understand is that the bubble you live in is really customized. Because what these algorithms are doing are learning. What do you engage with? Give him or her more of that. And so you're, you're sort of going down this rabbit hole and it really is quite bizarre. And I don't think, it's one of those things, you know, where you, you put a toad on a stove top and you turn up the heat very slowly, it doesn't notice. You can't notice and all of a sudden it's over, right? Your, your, your brain is melted. I think that's sort of what we've been doing for 20 years and I don't think people have been noticing. And I do, what worries me, Jeff, is that I think it's fine to have political disagreements. It's fine. We should, we should disagree on things.

• 7:20Speaker 00

  There are things to disagree about, but we can't disagree on the fact that two plus two is four. There are no opinions here. There's no, like things, either things happen or they didn't happen. And we are now arguing about facts. And that is a dangerous thing for a society, a democracy, and I would argue for an economy as well. Like we have to have a shared sense of reality. We can disagree on lots of things. I'm okay with disagreements. I disagree with people all the time. But I can't have people calling me up and saying two plus two is seven because I saw it on Facebook. And that is the tenor of the conversations that we are living in right now.

• 7:56Speaker 01

  It's a super scary prospect and, and you know, you use the word worry. And so if you, if you play the clock forward here, what's the implication? If we, if something, I guess dramatically different doesn't happen? What, what happens to our world if we continue down this path? And, you know, and as you said, it's not even necessarily a stagnant path. It's one that feels like it's marching in the wrong direction.

• 8:20Speaker 00

  Yeah. I, I don't think it is hyperbolic to say this. I think these are existential threats to societies, democracies and economies. I really do. And look, you're sort of seeing it right now play out. We are living in this country right now in two or three or four alternate realities. How the right sees facts, if, if you can call it that, and how the left sees facts. Again, if you can call it that. And by the way, this isn't a partisan issue. Both sides of the political spectrum have a problem with echo chambers. Let's be clear about that. But we, we can't have civil conversations because one side says, you know, X and the other side says, no. And that's it. That's sort of the end of the conversation.

• 9:05Speaker 00

  You know, my wife and I, my wife is a computational neuroscientist and she's incredibly smart. And one of the things she always does when we have a conversation, she said, let's start by agreeing on the facts. Because if you can't agree on basic facts, you're nowhere, right? There's nothing left at that point. And what's amazing about most of our conversations is sometimes we don't actually agree on, on the facts and then conversation's over, right? There's nothing, there's nothing to be said anymore. But I do think, I don't know how we have a democracy and elections when any side of the political aisle can say anything they want. And it doesn't matter if it's true or not, that sort of post-truth world. And I think that's dangerous. Here's another thing that's really dangerous. You know, it's easy to, to jump up and down on the heads of social media and, and the Mark Zuckerberg's of the world. It's easy.

• 9:50Speaker 00

  But the other thing you have to realize too, is that the large language models are in some ways more dangerous. And here's why, like, like if Elon Musk doesn't like something that his Grok AI is saying, he can put his finger on the scale and change it. In fact, that's not hypothetical. He has done that, right? We know that when Grok says something that is negative towards Elon Musk or somebody he likes, he changes the AI algorithms to not say that. And now imagine if we are starting to get our information not from social media. Great. But what if we now are getting all of our information from Grok and Claude and ChatGPT? And now there's three people in the world who control the flow of information because that's their AI.

• 10:34Speaker 00

  Right? Imagine it's a state-sponsored AI. So I think that's, you know, that, that's now almost the opposite of what the internet promised to do. It's reconsolidating the flow of information under the hands of a couple of billionaires, trillionaires. And I would argue that is really dangerous for our democracy and society.

• 10:54Speaker 01

  If you don't work in enterprise technology, you can skip ahead or take a quick braincation. But if you do, InfoTech is your secret weapon. InfoTech helps IT teams get projects done faster, better, and at a lower cost. Whether you're rolling out new tech, improving processes, or just looking for cost savings, InfoTech provides unlimited access to practical tools and expert guidance you need to execute at a fraction of the cost of traditional consulting. No matter the project, from AI strategy to cybersecurity to vendor negotiation, InfoTech has you covered. Check it out at the link below and don't forget to like and subscribe.

• 11:33Speaker 01

  Can we maybe take a step back and just kind of map the landscape a little bit more? Because I think if I, if, just to summarize, a lot of the message I've been hearing is that we are being manipulated through information at a mass scale by a number of different actors for a number of different purposes.

• 11:51Speaker 00

  Good.

• 11:52Speaker 01

  Can you map that landscape a little bit for me? You talked about it quickly.

• 11:55Speaker 00

  Yeah. So first, I think that's a good summary. You know, I'm a professor, so I rant and rave. And you're a professional, so you summarize it in 10 words. So well done. Yes, we are being manipulated. But let's talk about who and how. So first, the vast majority of Americans, in fact, the vast majority of world citizens get the majority of their information from social media today. That's a sad but true statement. Okay? Now, what is the business model of the Facebooks and the TikToks of the world? They give away their product for free and they want your time and your eyeballs to deliver you ads to get you to buy stuff that you probably don't need. That is a trillion dollar economy in a nutshell. It's the attention economy, right? And what these algorithms have done, they didn't set out to burn the place to the ground.

• 12:39Speaker 00

  But what they learned is deliver content that drives engagement. It's a very simple optimization problem. And they don't care what the content is. Good, bad, ugly, illegal, horrible, it doesn't matter. Whatever drives engagement, you deliver it. And so you're being manipulated algorithmically on a very individual level. Now, who's manipulating you? Well, lots of different people. First, the tech companies are manipulating you. That's the algorithms. You're being manipulated by state-sponsored actors. Russia, China, North Korea, Iran are manipulating us. They're sowing civil unrest because a destabilized United States is good for them. So they're state-sponsored, sophisticated actors. Cybercriminals are trying to separate you from your money, right? Frauds, scams.

• 13:23Speaker 00

  Here's a cure for cancer. I've taken your kid. You have to send me $2,000. I'm going to extort you with non-consensual intimate imagery. All forms of scams and frauds that are now being supercharged by AI. They were there, of course, before. And then there's the ideological. People who are just ideological. They have a particular view and they are trying to push that as hard as possible. And then some people, it's just they're trying to monetize. They have no ideology. They're individuals just being like, I'm going to create content in order to drive engagement so I can make a couple of bucks every month. And up and think about that scale, by the way, from state-sponsored actors to some guy halfway around the world who's just trying to create some content that delivers ads for him, that delivers some money for him, and everything in between.

• 14:08Speaker 00

  And here's the thing is the harm to you as an individual is wide-reaching. There are absolutely these frauds and scams, everything from ransomware and malware and extortion, getting you to buy products that are going to be damaging to you or at best not helpful. People trying to separate you from your money. There are dangerous cybercriminals who are targeting your children, extorting them and trying to get them to do horrible things and leading to horrific consequences. And then there are state-sponsored actors that are sowing that civil unrest and everything in between. So it's our mental health. It's our physical health. It's our dollar. It's our democracy. And everybody's got a different sort of goal here. Some of it is ideological and some of it is monetary.

• 14:52Speaker 00

  And meanwhile, the platforms are, have become trillion-dollar platforms because they're indifferent to it, of course.

• 15:01Speaker 01

  It's a scary landscape.

• 15:04Speaker 00

  I like the slight pause there, by the way. It was really, it was sort of an appropriate pause because how do you come back from that?

• 15:09Speaker 01

  No, well, exactly. I'm just like taking it all in and just thinking to myself, wow, that's a, that's a bad scene. So, you know, not to get ahead of ourselves, but you know, the obvious next question to me is like, what do we do about that? And let's start, let's maybe start as individuals. I guess there's the smaller question of, you know, if every second piece of media we see is fake, can we get better at identifying that? Or is there some sort of broader behavioral change needed here?

• 15:41Speaker 00

  Good. So I think that's the right question to ask. So here's, here's the thing you have to understand. You cannot, as an individual, become a digital forensics expert. You are not going to be able to tell the difference between real and fake. And that's not just me mouthing off, being arrogant. We do studies in my lab. We do perceptual studies where we show people images. We show them, we have them listen to audio. We have them watch videos in a very controlled lab environment that is not emotional. It is not designed to trigger emotional responses. We are setting you up for success and we're telling you, half the content you will see is real, half the content is fake. Do your best. And people are basically slightly above chance. And that's in a controlled, unemotional lab environment.

• 16:24Speaker 00

  Put them out in the wild, you know, doom scrolling. You are not going to do this. It's just, it's over, right? Your visual system and your auditory system have failed you. You are not going to do this. And if you, if your listeners take nothing else away from this, take this away. You cannot do it reliably. I do this for a living and I've been doing it for a long time. It is really, really, really hard. And the worst thing you can do is think that you can do this well. That makes you more dangerous. I call it the arrogance ignorance problem. Very dangerous. It's fine not to know how to do something as long as you know you don't know how to do it. So I don't think we can protect consumers at that level. And it's unfair. It would be like saying, yeah, when you go to the grocery store, we're going to teach you how to figure out

• 17:09Speaker 00

  if the food is going to kill you or not. We would never accept that. We'd never say it's my responsibility every time I get in my car to make sure the brakes don't fail. No, no, that's the car manufacturer's problem. So I don't think we should put this on the consumer. Now, who should we put it on? I think there is, there's really sort of two levers here. You have regulatory and liability. Regulatory says we're going to pass laws that basically force the tech companies to do better. And we've been talking about this for 20 years. And honestly, you know, here in the US, we are almost nowhere. The EU, the UK, Australia, Canada, they're doing a little bit better. But this is really hard to pass laws. The laws move at a snail's pace. The technology moves 10 times faster.

• 17:51Speaker 00

  Lobbying efforts are phenomenally sophisticated and effective. I'm not that excited about the, we should, we should try. We should try to put guardrails on this. But if you ask me where would I, I think is probably the best place now, I think it's with liability. And I don't think that's a great answer, by the way, but here's why I think it will work. The reason why the physical products that we buy and bring into our homes are so safe is we created product liability for these companies. We said, you create a product that you knew or should have known as caused harm, we're going to sue you back to the dark ages. And it turns out when you do that, which is an imperfect system, we should recognize. But when you do that, the companies internalize the liability. And they say, okay, if we don't make a safe product, this is going to be bad for us. Let's make a safe product.

• 18:36Speaker 00

  But we haven't done that with the tech industry. We let them off the hook for 25 years. I think that's starting to change. It's changing with the, with the ruling in New Mexico recently, the ruling in California. The EU is cracking down, saying, look, guys, this isn't fun and games anymore. It's not like Las Vegas. What happens on the internet stays on the internet. This isn't, this is having real world consequences and people are dying and people are being hurt and people are losing money. 10% of Facebook's profits come from frauds and scams. 10%? That's insane. If that was true of a bank, they would be shut down immediately. All right? So we have to figure out how to say to these tech companies, look, guys, you have to create safer products.

• 19:21Speaker 00

  You cannot internalize the profits and externalize the risks and harms. Party's over. And that would be my first sort of go-to. On the behavioral side, I would say there is something people can do. You're not going to become a digital forensics expert, but here's what you can do. Get off of social media. And I know that's a sort of a, sort of a dumb thing to say, but it is sort of, everybody, even people who are on it, know it's the right thing to do. Here's what I tell my students, by the way, that works. I start there. I say, look, delete everything. You will thank me later. I swear to God, your mental health, your physical health, and I swear to God, your IQ will probably jump 30 points. But I recognize that by design, these products are addictive. So what I tell my students, and this actually works, is just take it off of your phone.

• 20:06Speaker 00

  You can keep the accounts, it's fine, I don't care, right? But take the apps off of your phone. And the reason you do that is that every spare minute you have of your life, you're not doom scrolling on these devices. And if you need to have the accounts, when you're back at your desk, you can go look at them. And I swear to God, students, you know, they have exactly the same pattern when they do this. They do it, they say it was terrible for the first week, just like quitting cigarettes or coffee. And then after that, they're like, oh, my God. And then they basically just stop going. That's the thing you have to understand. Look, he's deleting the apps off his phone. I can see him right now.

• 20:42Speaker 01

  I'm doing it right now. I'm doing it right now. I'm like, when you say keep it on the desktop? When you say keep it on the desktop, I'm like, I can do that.

• 20:50Speaker 00

  Yeah.

• 20:50Speaker 01

  So Instagram is gone.

• 20:52Speaker 00

  Yeah. You have to understand, these things are addictive by nature. And they are terrible. By the way, people can't see it, but I swear to God, he's deleting the apps off of his phone as we speak.

• 21:02Speaker 01

  Reddit, goodbye.

• 21:03Speaker 00

  One more down. I'm doing this one at a time.

• 21:06Speaker 01

  One person at a time. I'm gonna be an evangelist for the message, too. I love it. I love it.

• 21:10Speaker 00

  It's like a pyramid scheme.

• 21:12Speaker 01

  It's the anti-pyramid scheme, right? It's deprogramming. It's un-brainwashing. No, that's awesome. Sorry, I got distracted deleting my apps. But I'm very happy, that's a very happy distraction. So, you know, we kind of covered both sides there, which I appreciate. I guess sort of the demand for the social media stuff, the supply of it. Coming back to the supply, and of course, correct me if I have my facts wrong, but my recollection is, there's a case going on right now, I want to say in Texas, against, you know, Meta and Instagram. How much liability there is for creating these things that are deliberately addictive here. So I'm going to assume, based on what you said, that you're an advocate for cases like this.

• 21:55Speaker 01

  You think that there should be ramifications for that. So, I mean, first of all, let's unpack this. Is that the right approach? And assuming this goes perfectly or close to it, what changes would you like to see big tech make that make this, you know, a more, you know, controlled and low harm environment for consumers?

• 22:17Speaker 00

  Yeah, so I would say I'm a reluctant advocate. I don't think liability is really the best solution, but we've tried everything else. We've tried regulation, we've tried boycotts, we've tried asking, begging, pleading, threatening. Nothing works. And so this is sort of where we have landed, and I think it will affect change. I think it's an imperfect solution, but I do think it's sort of the thing that we have that seems to be effective. Number two is, I'm not saying, I just want people who are victims to have their day in court. This is not for me and you to decide. It's for a jury and a judge to decide. And let's just have the conversation, right? And if the judge and jury say, hey, look, we don't think Facebook did anything wrong,

• 23:02Speaker 00

  I can live with that verdict, but let's have the conversation. Let's have our day in court. And we have not had that for many, many, many years, because the tech companies have managed to avoid liability for that. So I think this is good. Now, in terms of the changes, here's what's great about the liability solution is, I don't have to tell you what to change. You're gonna figure it out for yourself. Here's the thing that makes me crazy about the tech companies. They have built phenomenally complex systems. And yet, every time, every time, it comes down to, how do you make your product safer for kids? How do you make sure scammers and terrorists are not exploiting your platform? They're like, oh, man, that's a really hard problem. We don't know.

• 23:45Speaker 00

  I'm like, no, you're a big fat liar. Because every time you need to solve a problem that you can profit from, you seem to have no problem whatsoever, by the way. And yet, when it comes to the bare minimum to keep people safe, you play dumb. I'm not falling for that anymore. So the companies, this is what's great about liability is, you don't want the US government telling the companies how to make their products safer. It's not really what you want, because the government doesn't know, and the products move too fast. The companies are gonna internalize this liability. They're gonna keep writing 100, 200, 500 million dollar checks. And they're eventually be like, all right, this isn't worth it anymore, right? The scales have tipped, and they're gonna start making their products safer. Here's a few things I would advocate for, though.

• 24:28Speaker 00

  Australia just passed a ban on social media for kids under the age of 16. I think that's awesome. I'd make it like 30, but I'll live with 16, right? I think that's fine.

• 24:37Speaker 01

  And over 60 while you're at it, too.

• 24:39Speaker 00

  Oh, and over 60, yeah, yeah. It's a fair, that's a really good point, Jeff. Yeah, you don't want, you want your boomer parents on it. I think that's good. I think we've been doing an experiment on young kids, and I think it's overwhelming evidence. This is very bad for them. It's bad for them. So I would say, really force these companies to make sure they know who their users are. And don't let them play dumb, by the way. TikTok said, somebody, some C-suite in TikTok said, we know just about everything about our users within the first 90, 120 seconds. They know everything about you. That's the whole point of the algorithms, right? That's what they tell their advertisers.

• 25:21Speaker 00

  Oh, you want to target this very tiny little, they know how old you are, they know where you live, they know everything about you. So they can do this, right? They have got to reinvest in trust and safety. They've been eviscerated over, these teams have been eviscerated over the last few years. And here's the big one, and I don't know how to do this, is, I think the original sin of social media is the business model, right? When you give away your product for free, we as the users are not the customers. We're the product, right? We're, it's the wrong business model. And the reason we have this business model for the young people listening is, 25 years ago, I know this is gonna sound crazy, is nobody thought you can make money on the internet, which is sort of hilarious now that you think about it.

• 26:04Speaker 00

  But in the early days, remember this, Google, Facebook, it was like, guys, this is really dumb. You're never gonna make any money. Turns out, the joke was on them. So I think we have to really fundamentally rethink the business model in a really deep way. And one way that we can do that is, this is where the government can step in. And, because one of the things that these tech companies have become is just gigantic monopolies. And they can squash anybody in the competitive landscape. Let the market be sort of open, and let's make room for new business models. We live in a different time right now. And, and the problem is, every time there's like a fresh new startup with a new idea, one of these big tech giants comes here and just wipes them off the face of the earth, right?

• 26:46Speaker 00

  And we need, we need sort of a healthier ecosystem so that better and new ideas can come around. The age of Mark Zuckerberg is over. We need, we need, we need the next better version of that. And so I think that's where the government can come in and create incentives for new companies with better models, and protect them from the big giants. And by the way, I will remind everybody that in the early days, when Microsoft dominated the landscape, Google was trying to come up in the early days. And they went and complained to the Department of Justice, saying, Microsoft is destroying us with their anti-competitive monopolistic patterns. And they were. And what did the DOJ did? They went to Microsoft and said, knock it out, knock it off. And now, Google is saying, no, no, government regulation, bad.

• 27:31Speaker 00

  Government, bad. Guys, you are here because the government protected you. You can't walk through the door and then close it for the next guy, right? So let's stop pretending somehow that the government is inherently bad. The government can do things to create fair, open ecosystems where people can compete.

• 27:49Speaker 01

  So here's where my mind goes and what concerns me. You talk about needing that next generation of kind of tools and products and platforms. And it's looking more and more like we have that, and it's LLMs, right? It is these AI, generative AI tools. And as you said earlier, in some ways, they're even worse than the incumbents, right? Because, you know, you've got the Elons of the world who can put their thumb on the scale, and it may be imperceptible to us in ways that are more subtle, if we're unlucky, than, you know, the current roster. So what do we do about that? When it comes to, you know, the generative AI tools that people are using more and more, and, you know, if I had a crystal ball, I'd say are gonna be using more at the end of the year than the start of the year, and more next year than this year, what do we do there?

• 28:34Speaker 00

  Yeah, first of all, it's a great question. So the danger here, just to put a really fine point on this, is, the companies that have built like the best LLMs, large language models of the world, are sort of the winners from the last two decades, because they have all the data, and they have all the money, they have all the compute. And, you know, some tiny little startup, it's hard. It's not impossible, right? You're seeing some movement out of these small companies out of China, although there's some controversy as to what they're really doing. But yeah, I worry that the winners from the last 20 years are gonna be the winners of the next 20 years, and how are we going to compete? I think there is a potential solution here, which is that we have, the government has to step in and say, look, either we're going to create a fully open source model

• 29:16Speaker 00

  that is ethically trained, open source weights, anybody can tap into it and build the next generation. The government should do this. It's awesome. Because if you allow anybody to step in with an open source model and use that as the base, imagine the ecosystem that you can create. And we're not relying on Sam Altman and Elon Musk and a handful of companies who are the last year's winners. I think there are things that the government can do to create a level playing field here. I don't know what the future of large language models looks like, but it's sure looking like it's gonna play, it or its, you know, variations on it are gonna play a huge role everywhere. And, you know, in the same way that access to the internet is essentially now the same as access to electricity or water,

• 30:00Speaker 00

  probably access to this kind of agentic AI, LLM, whatever you want to call it, is going to be of that equivalent. And we've got to figure out how to make sure that that is not consolidated power, both for the fairness of the ecosystem, but also, as we were just talking about, you don't want three people in the world who can put their finger on the scale and change all of the information that everybody sees instantaneously.

• 30:25Speaker 01

  Honey, listening to that, it's really interesting and it makes complete sense about, you know, having the government play some sort of role here, trying to get to more of a true open source model. And, you know, one of the reactions that I had to that is it's funny and I guess ironic, given, you know, the, the topic of this conversation, that I feel like there's such a strong narrative these days about like, that's the last thing you should do. We have no choice but to let these companies do it themselves. They're the only ones who can do it. You have to put faith in them. The reason I say that's ironic is because to me, that's a narrative that's so obviously being shaped and manipulated by the companies themselves.

• 31:01Speaker 00

  Sure.

• 31:01Speaker 01

  And there's so much, you know, whether you call it misinformation or disinformation, there's so much narrative shaping by, you know, big players, invisible players here. What do you consider to be, I guess, the most harmful narratives right now, or the most BS you're hearing in this entire space?

• 31:20Speaker 00

  I'll tell you, I'll tell you, I'm glad you asked this question. So first, and I hear this all the time, is from the Sam Altman's, the Elon Musks, all the big AI and tech companies, do not regulate us, do not create liability because China will beat us. That's the narrative, right? This is a national security concern. And the reason why this is such a BS is that creating safe products is a national safety concern, right? Of course it is, right? It would be like saying, look, guys, we need to create the next airline that's gonna compete, a next jet that's gonna compete. Don't force us to make it safe because then

• 32:01Speaker 00

  we'll have to compete with China. Yeah, I'm not flying on that jet. If, if the motto of the airline is move fast and break things, nobody's flying that airline. And yet, here we are in Silicon Valley, you know, everybody's like, oh yeah, sure, move fast and break things. No safety, no precautions. This is a national security concern if we build unsafe products. And by the way, we unleash them in critical infrastructures, making decisions about who gets targeted in war, controlling our energy infrastructure. You don't think that's a national security concern? Are you out of your mind? Don't let them get away with this. This is not an us versus them. This is a convenient line. And by the way, this is the same BS we heard in the last 25 years. Get out of the way of the internet,

• 32:44Speaker 00

  let the internet be the internet. You're going to stifle innovation. And the reason why you should reject that is we do not accept that in any other industry. The car industry doesn't get to say, get out of the way or we won't be able to compete with China. No, we say, make your products safe. Test them. Make sure I'm not gonna explode when I get into my car. Medicine, food, like we force safety and it turns out, that's good for consumers, it's good for the industry, it's good for everybody. And if you don't like the way another country is doing what they're doing, we have other mechanisms to deal with that, right? We don't, we don't say we are going to reduce ourselves to the lowest common

• 33:23Speaker 00

  denominator and tear the place to the ground just because somebody else is doing that.

• 33:30Speaker 01

  I think that's exceptionally well said, and I don't have a lot to add on to that. I really, really like that framing device. So maybe I'll switch gears a little bit and, and honey, return to something we talked about earlier, which is, you know, the detection of, of deep fakes or, you know, altered media. You know, you mentioned that, for the, like, for the most part, we can't do it. Like, we are screwed out here, you know, in the masses. You know, for the, for the geeks out there, how do, how do you do it?

• 33:59Speaker 00

  Yeah.

• 33:59Speaker 01

  If, you know, you're out there on the frontier, what, what does that look like to it, to like a true professional?

• 34:04Speaker 00

  Yeah. So let me start by saying, there's a little tension in this question, right? Because the more I tell you and your listeners how I do this, the more my adversary learns from what we do, right? And there's always been this tension in cyber of, you know, how much do you give up? How much do you not give up? But having said that, let me tell you a few things. So first, there is no, there's no like just magic button. You don't put a button on the keyboard and say enhance, right? And then, or, you know, analyze. That is just not how it works, right? Every image, every audio, every video is just different. So let me give you a couple of examples. So here, here's one I really like. So for example, this was just yesterday, we were dealing with a piece of content from a major

• 34:49Speaker 00

  news outlet, and it was a video taken outdoors of an explosion. Yeah. And we really like outdoor images and videos. And the reason we like them is that we actually understand them fairly well. So we know, for example, that outdoors, there is one dominant light source, the sun, right? You can see it on my background here. And the sun has some very particular behaviors, right? Like all the shadows in the scene will be consistent with one light source. The pattern of illumination will be consistent with one light source. And so we can physically measure, what is the pattern of illumination in the scene? And is it physically plausible and consistent across it? So we have a, a light pole over here, we have a person standing over here, we have a building over there. Is the pattern

• 35:34Speaker 00

  of lighting and illumination consistent across that, the scene, right? Physics based, explainable, not give it, you know, tell, ask the AI what this is. That doesn't work. Don't, first of all, here's another PSA. Do not, do not, do not. Upload a piece of content to some website you found that says that it has a 99% accuracy in detecting real versus fake. They're lying, it doesn't, and it doesn't work. Those things don't work. They're all nonsense. By the way, I found three of those sites that I swear to God were random number generators. They're just, they're just stealing your data. They're just, your data, and then they're just giving you a random answer. And people are like, oh, it's so smart. And you can just go look at the JavaScript code. It's actually not

• 36:17Speaker 00

  that hard. Okay. So physics based modeling of the physics of the world, the, the lighting. When it comes to people, we have a whole different set of techniques. So here's one of my favorite ones. One of the, one of the things we've been seeing that's really problematic is on calls like this, by the way, you're talking to somebody and they can change their identity in real time. They can do what's called a face swap deep fake from eyebrow to chin, cheek to cheek. You can just change your face. And we've seen, for example, North Korean IT workers penetrating Fortune 500 companies, cyber criminals coming in and plant, getting, they're, they're not going through the firewall, they're actually going through HR, getting jobs, and then they're on the inside stealing IP and planting. Huge

• 37:00Speaker 00

  problem across the industry, by the way, huge. So how do they do this? So what they do is they take your face, for example, a single image of you, and then as I'm standing in front of the camera, it's just putting your face on my face, no matter how I move and how I blink my eyes and talk. And it's like, it's called a face swap deep fake. Here's what's really cool about that. It's, think about, like, if I put on a rubber mask, as I move, the mask conforms to my face. Well, that's what you have to do digitally with a face swap. And when you do that, it's constantly like shifting the pixels, right? It's called warping. And warping has a very specific mathematical definition. And so, and we, we can define it. And then when we see a video like this one, either in a stream

• 37:45Speaker 00

  or in a, in a digital file that we saw on social media, we can detect those very specific manipulations. Yep. So a whole host of these techniques that we've developed. But let me, I want to just pop up one level here and say, the way that I don't think you should do this, and the way a lot of people are thinking about this problem is, this is an AI problem, you solve it using AI. Get a bunch of data, throw it into your favorite transformer, deep neural network, whatever it is, and have it learn. And the problem with those approaches is twofold. One is, they don't really work. And you don't know when they don't work. But here's the bigger problem. You can't explain what they're doing. You get to the other end of it, and imagine you're in a court of law, and you're somebody

• 38:28Speaker 00

  like me, and say, hey, we've done a forensic analysis of this piece of content, and it's fake. And the judge or the prosecutor says, well, how do you know? And I gotta be like, well, you know, the computer told me so. I mean, that's not an answer, right? I mean, if you're gonna accuse somebody of, of, of fraud, or committing a crime, or manipulating media, you gotta be able to say something. So we are driven both academically and over at Get Real Security, really understanding what this problem is, not going into it blind and just hoping for the best. And so we have physics based techniques, we have geometric techniques, we have statistical techniques. We are not antagonistic to AI. I don't mind bringing AI into the, into the loop when we need it, but it's, it's very judicious,

• 39:10Speaker 00

  if you will. And then, you know, there's a lot of other stuff I won't talk about. Really cool stuff, things that, you know, we don't want necessarily to be known because it's adversarial, right? What we're doing is adversarial. But a lot of it is really just understanding the 3D physical world, understanding what we expect to see in a natural video, image, audio recording, and then looking for deviations of that that are representative of AI.

• 39:39Speaker 01

  It's very cool. And as you were speaking, it just reminds me that, yeah, I'm probably not going to be able to do that. And most people aren't as we're kind of idly scrolling through our phones. You brought up a scenario that we haven't talked about yet that I think warrants further exploration, which is like the enterprise scenario, the North Korean hacker that's made their way in. And by the way, there's the versions of this too, where you get a call from the CEO and they say, I need a wire transfer, whatever. There's versions of this. What, if you're speaking with business leaders, with technology leaders, what are the enterprise scenarios that they need to have on their radar in this space?

• 40:20Speaker 01

  And what are the precautionary steps you would recommend there?

• 40:25Speaker 00

  Good, yeah. I'm glad you talked about that because we have been talking about consumers, but let's now talk about the enterprise. Let's talk about trillion dollar economy. So first of all, these are not hypothetical threats. When I say North Koreans are, this is a multi-billion dollar a year industry for North Korea. It's how they're funding their government is with IT workers who are working in US. I mean, it's, you can't make this stuff up. You mentioned the CEO attacks. These are not made up. 10 million, 20 million, $30 million scams. And by the way, for every one of them you read about in the newspaper, there's 10 that you haven't read about. Brand damage, people creating videos of your CEO or your brand, getting them online because they don't like you. They don't like something you said.

• 41:09Speaker 00

  The damage that can happen to a brand is unbelievable how fast it can happen. So here's the thing that enterprise has to understand. This is not an if, this is a when. And I can tell you, I, over the last four years at Get Real, I have talked to a lot of C-suite, a lot of boards, a lot of CISOs, a lot of CEOs. You either have been attacked and don't know about it, or you've been attacked and you know about it, and it's coming. So let's go down the list of what we are seeing. So first, so much of our day-to-day interactions are now on these things, right? And they're great, they're convenient, right? If you want to interview somebody and they're 3,000 miles away, you don't have to fly them out. Just get on a call and talk to them, right?

• 41:49Speaker 00

  But it's also a vulnerability because we know cyber criminals bait and switch interview. And my favorite one is people hire other people to do the tech interviews. That person goes through the process, gets an offer, and then two months later, somebody else shows up to work and nobody's connecting the dots. Right? It's two months later, like, yeah, we interviewed this guy. Okay, sort of looks like it. My God, who can remember? So you're seeing that, you're seeing the North Korean imposters. You're seeing people who do what's called overemployment. They have five, six jobs and then they just outsource it to India. And then they're just in, you know, it's completely crazy. And by the way, a lot of those companies are competitors of each other. So there's this weird crosstalk between these things.

• 42:30Speaker 00

  So absolutely onboarding from the interview process to the onboarding is a huge vulnerability. And I think enterprise is starting to wake up to that. You're starting to see that. There is fraud, particularly in the financial sector. So if your bank, for example, uses my work, my voice is my password, you should change banks. Your voice is no longer secure. With 10 seconds of your voice, Jeff, I can clone it and I can call your bank and I'll pass any biometric. So fraud at the financial institutions has become a huge problem. Insurance companies. Somebody smashed into my motorcycle a couple of months ago, totally trashed it and drove away and called the insurance company. Oh yeah, no problem.

• 43:13Speaker 00

  Take a video of it, send us the video, they write me a check. I'm like, well, that was cool. I'm like, do you guys not know how AI works?

• 43:23Speaker 00

  I didn't actually say that because I wanted them to fix my bike.

• 43:25Speaker 01

  You wanted the check, yeah.

• 43:26Speaker 00

  So insurance companies are going to deal with this problem. Airbnbs, right? People are, real estate, people are creating fake images of houses, Airbnb's, Ebays. Anything where commerce is happening online has become a problem. The brand protection I already talked about, up and down the stack, you're seeing the enterprise being attacked. And my sense of the landscape is some of the enterprises waking up, they got whacked. This is what happens in cyber, right? Everybody's like, oh yeah, it's a problem. And then they get hit and they get hard and now they have to respond. And my thing is, guys, get out ahead of this. You know this is coming. You know it's coming. Everybody knows it's coming.

• 44:09Speaker 00

  But of course, you know, they're dealing with 100 different threats. So I think that we are going to see more and more problems in the coming, I was going to say years, but really months. And you're seeing it escalate, we're seeing it escalate. Just yesterday, I was talking to the president of a large bank and he's getting attacked. People are creating fakes of him, circulating them, and it's bad for his bank. It's bad for him and it's happening every day.

• 44:37Speaker 00

  So, and here's the thing is, anybody can do this. I mean, the reason why this is happening is a single individual can do this with a keyboard and access to internet. It's nothing, there's no barrier to entry anymore. And so that's why I think the threat is becoming so sophisticated, so much more prevalent.

• 44:55Speaker 01

  So let's get, let's get super tactical about this. You talked about organizations waking up to this, you talk about getting out ahead of it.

• 45:02Speaker 00

  Yeah.

• 45:02Speaker 01

  What does that look like? If you're one of the lucky few who says, yes, this is a problem, I know it's a problem. I'm worried about onboarding, I'm worried about deep fakes. I know that there's so many employee-centric vulnerabilities here. What does that look like? Is it awareness, training? Like what types of tactics are going to be, I guess, the highest impact in terms of protecting your organization?

• 45:28Speaker 00

  Yeah, I think there's three things every organization should be doing right now. And it's got to start with, as you just said, Jeff, awareness. You gotta know. And look, nobody wants to do those silly trainings on cybersecurity and all these things that we do. But it turns out, you know, they're actually pretty helpful. This just keeps things at top of mind. Know what a phishing scam is, know, you know, remember that, oh, don't click on the link, for God's sakes, don't click on the link. I mean, so that, that kind of awareness is good. And a lot of people don't know, for example, that on a real-time call like this, I can change my voice and my face. And my, and my, I'm sorry, my voice and my face. They don't know. And just not knowing is already a vulnerability. So I think awareness training is really important. Two is policy. You gotta start creating some policies.

• 46:09Speaker 00

  I mean, I've talked to some companies who said, all right, no more remote interviews. I think that's insane. I think that's an insane response to a threat because it's so extreme in some ways. But you know, you gotta start putting in some policies that say, look, when people join, how many calls have you and I been on where people join from some phone number and nobody knows who it is? And their camera's not on? How are you in a business meeting? It'd be like somebody walking into your business room and nobody knows who they are. And everybody's like, ah, it's fine. Like, we gotta start having some more reasonable policies. We have to stop being so cavalier about who's joining these calls from some random number without their camera on. So I think that there's just some good policy. And number three is tech.

• 46:53Speaker 00

  So this is, you know, obviously my bread and butter now. So for example, over at Get Real, what we have done is we have developed technology that can join a Zoom call, a Teams call, a WebEx call, any, any video call. And it monitors. It's, think about it as these, the transcribers, the things that are transcribing your calls, right? Sort of annoying, but pretty cool actually at the end of it, right? You get these nice summarizations, you get action items, who's doing what, really, really good. So it's the same type of thing. It's a little, what we call trust advisor that joins your call and it monitors. It monitors the face and the voice of all the participants. And then you get to decide at the back end, well, what do you do if you find a fake? Right? Do you notify everybody? That's probably not a great idea, right? Because the bad guy's on the call. Do you notify the moderator?

• 47:37Speaker 00

  Do you notify somebody in the, in the chief information security officer? Like you have to decide on policy there. That's where the policy kicks in again. I think if you're interviewing people remotely, you're out of your mind if you're not using this technology. You are absolutely out of your mind. Because I can tell you, you're interviewing people who are not who they say they are. So, and it's a pretty light lift, honestly. You're, you still get to use Zoom, you're not flying people out, which would be an insane response to this or disproportionate, I would say. So I think there are technological solutions here that will help. I think the combination of training, policy and tech will start to protect institutions way more than most of them are now. I'll add one last thing to this, by the way, this is not hypothetical.

• 48:21Speaker 00

  We know that the North Koreans have penetrated Fortune 500. We talked to one company who I won't mention by name. It's one of the largest employers in the world. They caught 1,800 potential North Koreans. That's not a typo, 1,800, right? And that's the ones they caught. Now you gotta ask yourselves, how many got through? So these are not hypothetical threats. These are, these are happening to your organization. We talked to one US defense contractor who had five North Koreans in their organization, five. That's insane, right? So I don't think there's a single Fortune 5 company that probably hasn't gotten hit, almost certainly. And it's just a question of whether they figure it out or not. So I do think there has to be audits.

• 49:06Speaker 00

  We worked with one customer that just didn't audit. They're like, okay, we almost certainly missed it. Go back and re-audit everything, right? So you gotta backfill it too. You can't, it can't just be forward looking. You gotta go back and backfill and figure out who are all your remote workers.

• 49:23Speaker 01

  It's extremely scary and the numbers are, I've used this word before a few times, they're staggering and they're like sobering in a pretty big way as well.

• 49:33Speaker 00

  I agree.

• 49:34Speaker 01

  Yeah, I'm glad to hear it.

• 49:36Speaker 00

  When this person told me 1,800, I'm like, 18? He's like, no, no, 1,800.

• 49:42Speaker 01

  Yeah.

• 49:45Speaker 01

  So I want to come back to something you said, though, you know, just to put a slightly different lens on it. So, you know, you obviously spend a decent amount of your time speaking with, as you said, CEOs, CISOs, CIOs, CTOs, who are, you know, in some way or another, being kept up at night by these challenges. What are the predominant themes or, I guess, actual activities that are on their mind? And what's kind of your-- what are the macro level themes of advice you're giving them?

• 50:17Speaker 00

  Yeah. First thing is, you enumerated a bunch of different C-suite, right? CEO, CISO, CTO, CIO. First thing is, nobody knows who owns this problem. And this is a problem. We were in a Fortune 500 company board meeting with a bunch of their C-suite. And somebody in the board said, this isn't a problem. I'm like, all right. And then five hands went up from five different divisions being like, no, no, this is what happened. But nobody is comparing notes because nobody owns this. Right? The CISO owns, all right, malware attacks, ransomware attacks, they own that. So nobody owns this. And so everybody's just-- so that's a part of the problem. And I don't know if we need a chief AI officer. I mean, God forbid.

• 51:03Speaker 00

  What a nightmare. But somebody's got to own this in the organization. You got to figure out who that is. And it's complicated because in some cases, it's more on the HR side. Sometimes it is more on the CIO side. Sometimes it is more CISO. And so you got to figure out who owns this problem. And there has to be better communication. After that, it really depends on the sector. Financial sector is dealing with things that are entirely different. You know, huge employers, you know, the Amazons of the world, dealing with something entirely different because, you know, they're onboarding like 100 people a day. Right? The scale at which they're operating is completely different. So I would say it's fast moving right now. The threat is moving very fast. And my thing is, when-- I've noticed this is so interesting.

• 51:48Speaker 00

  Like, when I physically go to a company, I have to go through all kinds of security checks. Right? I go to the desk, they look at my ID, they take a picture of me, they know exactly who I am. I have a badge, I'm allowed in certain places, not in other places. They have very tight controls in the physical space. When I get on a call with those same people from the same company, they're like, eh, get on the call. Anybody can-- I mean, it's just like this weird juxtaposition of the way we treat physical security and digital security. And I don't think those are fundamentally different. I can do just as much damage from here in my office on this screen as I can do in a physical space. So I think we have to start thinking about these interactions, these digital interactions differently. Here's another one, too, by the way,

• 52:34Speaker 00

  that was really interesting conversation is-- it was the same president of the bank the other day. And he said, look, I'm worried that when I get on calls, I don't know who I'm talking to. I'm like, yeah. He said, but here's the other thing I'm worried about. I'm worried that other people won't trust me when I'm not on a call with them, which I thought was really interesting. Like, when I get on a call, I want you to be able to trust that it's me. And I thought that was a really sort of cool idea, too. And I hadn't thought about it that way. So I think this notion of trust is going to become really-- and you know what's interesting about this conversation, as I'm saying it aloud, is this is not that different than what we were talking about 50 minutes ago. Trust, right? How do you trust what you see online? How do you trust this digital interaction?

• 53:16Speaker 00

  And I think we're going to have to get a handle on this. Because otherwise, you know, everything is up for grabs now. I don't know how we do business. I don't know how we consume information. Everything is suspect at this point.

• 53:32Speaker 01

  I'm glad you brought that up in the, you know, kind of close the loop moment because it feels like a different flavor of the same challenge. And, you know, if you can't trust that what you're seeing, you know, on your phone, on social media is real, how do you know that your CEO is real? And the same-- I have the same concern about this, which is, are we on a trajectory for this to get better? Or are we on a trajectory for this to get worse? And, you know, just based on a lot of our conversation, I'm--

• 53:56Speaker 00

  Yeah.

• 53:57Speaker 01

  I'm a bit worried that it's actually going to get worse before it gets better. So, you know, Hani, you said that the answer is probably not just shut it all down remotely and you better be in the CEO-- in the room with the CEO if you're going to trust him or her. So that's not the answer. How do we start rebuilding this trust digitally? Is it having some sort of authenticator on the call? What does trust actually look like if we're going to get over the hump of this problem?

• 54:25Speaker 00

  Yeah, yeah, yeah. First, I think you're right. It's going to get worse before it gets better. And almost everything does, right? This has to get bad. People are going to get whacked. It's going to be ugly. And then we're going to hopefully crawl our way out of this. Because I do think that's probably the trajectory. And I think this will unfold over the next few years. And that's what we've been seeing. So what does it look like? First, let's acknowledge that nothing is perfect. Right? But we are far from perfect. We're not even close to being good. I think we are going to have to figure out how to have trusted communication, number one. Like, when I get on a call with my bank, when I get on a call with somebody that I'm doing business with, I'm talking to a potential investor, I've got to figure out how they can trust me and I can trust them. Because I think the day of, all right,

• 55:09Speaker 00

  I'm on a video call with somebody, I know who it is now, is over. And by the way, that's a relatively new phenomenon. I mean, that really is just in the last two-plus years. So I think we're going to have to figure that out. And that's going to have to be a combination of tech, policy, awareness, et cetera. There may become a situation where you're like, look, guys, this is too important. We're going to do it in person. I was just having this conversation with some of the folks over at Get Real. Like, maybe there are going to be certain times where you're like, look, we're not having this call. I mean, I can tell you, my world is really weird. And I get on calls all the time. I was on the call yesterday and I was almost sure it was a deep fake. And I spent the first 15 minutes just scrutinizing because I couldn't add our trusted advisor because it was somebody else's call.

• 55:54Speaker 00

  Like, we can't have this. It's, you know-- and here's the thing, this is really important, too, is that we can't just enter a world where we distrust everything. Right? That doesn't get us where we want either. Because then you get on calls and you're like freaked out and you start hanging up on people. And that's no good either. Right? So we've got to find that middle ground. And I think this is a place-- and here's the good news about the technology here is that it can actually work really, really effectively. One of the things that is working to our advantage in real-time calls like this is the adversary has to create a fake at roughly one every 30 seconds, 30th of a second. Right? That's the frame rate of the video. I can wait five seconds to figure out if it's fake or not. I have a lot of time. And you and I have been talking for almost an hour now.

• 56:39Speaker 00

  So imagine that call unfolding, not for seconds now, but for minutes. We'll figure it out, right? We have a lot of signals that we can look at. So I do think that technology-- this is a problem, particularly in real-time calls, because we are analyzing as it's unfolding. And that's quite different than something's been posted on social media, it's been up there for an hour and has 10 million views. By the time I see it, it's very hard to unring that bell. So that problem is much harder. I mean, we are working on it. But this one, I think, is actually tractable. And that's actually why we're so excited to be working on it.

• 57:15Speaker 01

  I'm glad you've veered into the good news camp, which I like. And I wanted to sort of--

• 57:20Speaker 00

  After an hour.

• 57:21Speaker 01

  Yeah, well, to try to end on a positive note after there's been so much doom and gloom. And rightfully so, by the way. I think everything you said is completely valid and should be on people's minds. But just to continue along that thread, what makes you optimistic about, you know, the future of this space and the future of tech?

• 57:39Speaker 00

  Yeah. So-- look, I've been doing this for a long time. I've lived through the personal computer revolution, the internet revolution, the mobile revolution, and now the AI revolution. And I don't think there is a time in the last 30 years that has been this exciting, really. I mean, something really amazing is happening. I was sitting down a couple of days ago with a colleague of mine here at Dartmouth College, who's a theoretical computer scientist slash mathematician. And we were talking about how AI is now proving theorems, open problems for 100 years that mathematicians haven't been able to solve. AI is solving them. It's amazing.

• 58:24Speaker 00

  These are really, really hard problems. Like, a bunch of really smart mathematicians tried to solve these and AI came around in an hour, figured it out. Holy crap. That's amazing, right? We now build entire end-to-end software products for prototyping in hours what would have taken us six months. It's incredible the things you can do with this technology. There's no question-- I don't know, you know, where LLMs will hit the wall. I don't know what comes after that. But something really amazing is happening with this technology. And I don't-- you know, there's a lot of, you know, naysayers. There's a lot of people saying that, you know, the other side of that-- something is happening. And I think it's really exciting. And I think it's an enabling technology.

• 59:10Speaker 00

  You don't need to have a degree in computer science to write code anymore. That's amazing. We are going to start solving open math problems. That's amazing. I just think that we should learn the lesson from the last 25 years that we have to figure out how to harness the power of this technology while keeping people safe. And I think it's the second part of that that we forgot for the first 20 years. We harness the power of technology, but we did less of a good job of keeping people safe. So I think we should learn from that mistake. And I think you can have both. I reject the idea that somehow creating safe products is antithetical to making, you know, leading and innovating.

• 59:55Speaker 00

  I think they can coexist. And in fact, I think it's better for everybody. There is real fatigue out there around AI and tech. People hate Silicon Valley. They hate these tech industries because they're sort of running over-- they're running everybody over. We're not caring about it. And that's not good for the industry. The industry has to understand that. This is not good for you. And so everybody benefits when you make a safer, better product that we can all use without burning the place to the ground along the way.

• 1:00:21Speaker 01

  Well, Hani, that is, in my view, fantastic note to end on. I completely agree with you. And I think it gives everybody who's listening a lot to think about. So I wanted to say a big thank you for joining the program. I think it's been really eye-opening and insightful.

• 1:00:35Speaker 00

  Thanks for having me.

• 1:00:37Speaker 01

  Most viewers don't know this, but Digital Disruption is developed by Infotech Research Group, a leading advisor to technology leaders around the world. If that's not you, you don't need to care. So skip ahead and enjoy our content. But if you are a technology leader, Infotech helps IT teams get projects done faster, better, and at a lower cost. Infotech provides unlimited access to practical tools and expert guidance you need to execute at a fraction of the cost of traditional consulting. No matter the project, from AI strategy to cybersecurity, to vendor negotiation, Infotech has you covered. Check it out at the link below. And don't forget to like and subscribe.