AI Generated Images Spread During Iran War

Welcome back. AI generated images of the Iran war are circulating online. For example, fake images of the UAE's Burj Khalifa on fire or of US soldiers captured in Iran are adding fuel to the flames of an already intense situation. Here to help us make sense is cyber security analyst and former CIA hacker Eric Cole. Eric, thank you so much for joining us tonight. How do we differentiate between what's real and what's AI generated in this war? The short answer is if they're generated with high quality equipment and by countries like Iran, you really can't tell the difference. Now, if they're lower quality ones, the things you want to look for is the mouth movement and the audio, because very often with the low end deep fakes, which is what we call these videos, the audio and the mouth are not completely lined up and there's a lot of jitter on the face. But the reality is the stuff we're seeing produced out there, unless you trust the source of where it came from, I visually looked at some of these deep fakes and unless I took out advanced equipment, there's no way you can tell the difference. Yeah, I mean, just looking at what we're seeing on screen here as you're talking, it's shocking how realistic these look and it's not that hard to generate them, right? No, not that I'm encouraging any of our listeners to do this, but if you want to create sort of low to medium quality, you can download software from the app store or the Android store. It'll take you about two or three hours. You need about 30 to 40 seconds of video and audio of the person you're trying to deep fake, but literally that's all it takes. And then if you want to get into the really advanced ones, like some of the ones we see on the screen now, that's probably going to take some higher end GPUs and about two or three days. But once again, we're talking the cost is under $5,000 to create these high quality fake videos. So it's not that everyone anywhere can generate these kinds of videos at any time. You do need a certain level of skill, some investment. For the higher end ones, but the problem we're seeing out there is they're just flooding it with bad, good, and high end just to create a ton of confusion. Like if you go in and look at the news now, you don't know what's true and what's not because there's just so many fake videos that that's really the goal of even some of the low quality ones. They're not trying to convince you. They're just trying to confuse you. Yeah. And it really is confusing. So, you know, how do we go online and not just distrust everything we see at this point? The big thing, and you'll like this is trust news sources like yourself, like go for the reliable news sources that we know are not going to give fake news and try to stay away from the one offs. Like a lot of people like following these accounts that you don't really know who they are or where they came from. One of the big things I recommend is start unfollowing untrustworthy people you don't know and just stick with trustworthy sources and then you'll have less noise and more accuracy. So when you go online, I mean, what kind of mindset do you bring to the table? Because you're not just on there just reading things and trusting whatever it is. Essentially, I'm paranoid, a little paranoid, it goes well in this day and age, and I don't trust anything. So when I go in and see anything, I immediately go back and start doing research and say, okay, is this true? And this is an area where AI can help you. Because if you tune your AI correctly, like my AI engines are tuned, so they know trustworthy and they know fake, so they can go in and give me reliability rating. So I go in and I give a story to my AI engine and what is the reliability or probability that this is fake? And it will actually say 10%, 80%, or 90%. So this is an area where AI can actually help you fight the AI-generated deepfake images. Do you have any suggestions for what kind of apps to use or AI testers, so to speak? The one that I find, and once again, no endorsement here, but the one that I find the best is Claude. It really does a great job of going in. So every morning I say, okay, can you go through all the different news sources, read out the anomalies, and give me reliable news that's covered by multiple sources with a probability rating of under 5% that it's a deepfake? And then it will come back and give me my curated news. So that's really a good way to figure out what's real and what's not, as opposed to going on social media and just getting bombarded by all the noise. So using AI to, I guess, combat fake images and so on. Are these AI fakes having a significant impact on Operation Epicure, do you think? Or do you think that people can tell what's what for the most part? Has the public gotten used to seeing AI fakes or not so much? No, I think it's having a pretty big disruption. And what it's doing is it's adding what we call the psychops, the psychological component to this war, where now you're going in. And at times, I know people are like, oh, the war's not going well. Iran went in and eliminated these key leaders and those key leaders. And the reality is, it was all fake news. So it's really causing a lot of confusion, distrust, and a lot of political uprest of people getting upset and frustrated based on what they think is happening, when in reality, it's all fake and reality is quite different. Such a bizarre time we're living in. Social Media Platform X, of course, has a new policy that bans users from profiting off of unlabeled fake war content. What does that tell us about how widespread this problem is? It's huge. And a great example is if you look back a few weeks ago, where Anthropx, who actually is the company behind Claude, they actually got into a dispute with Department of Defense, where Department of Defense said, we're not using you because you won't give us permission to use your AI to spy on American citizens and to be used in weapons of mass destruction. If DoD is pushing those policies, this is a widespread problem that's not being used just with individuals, but also at the government at many levels. How do you think social media platforms should handle fake war images? To me, we need to have laws. The United States is one of the few countries that doesn't have unified federal laws on data privacy, on ethics and morals and social media. If you go to a lot of other countries, it's illegal to go in and do deepfakes and they actually prosecute and go after those individuals. In the United States, there's really nothing that makes it illegal for you doing a deepfake. And therefore, it's really hard for the platforms to enforce. And the unfortunate reality is the platforms make more money with deepfakes. So they're sort of in a situation where they're not representing our best interest, because if they went in and started doing filtering and content and made sure it was accurate, it would not only be very expensive for them, but they would actually lose money. So it's actually one of those things where we as citizens in the United States are in a bad position, because the government and big tech are not protecting us. Because there's a sort of an incentivization to push out this kind of content. It's something for X to actually say that they would de-platform people who do this. How significant is that policy measure that they put into place? It's once again, a move in the right direction. The key thing now is going to be enforcement. Are they really going to enforce it? And so far, from what I've seen, they're actually doing a really good job enforcing it and blocking accounts. So if they continue doing that, I hope they set the model for other social media platforms to follow and start really protecting and making sure the information we're getting is accurate and correct. Eric, before we wrap up, what concerns you most in regards to the use of artificial intelligence to create fake images and videos about the Operation Epic Fury? My biggest concern is that they're really trying to target the listeners, average citizens, people who don't think they're a target. Because most people when they think cyber attacks and cyber war and everything, they're thinking that they're going to hit big tech, they're going to go after large companies. But the reality is with AI, they can now mobilize and go after citizens. So I am seeing bank accounts being wiped out, crypto wallets being stolen. We're seeing a lot of attacks against individuals that have increased dramatically during this war effort that are not being covered because they're one-off instances. So anyone listening needs to recognize if you get any email or any text about a shipment or a toll, don't click on the link. Go to the app, apps are safe, links are not. And if you could just institute that one habit of not clicking on links and believing these videos, you can go a long way to protect yourself and your family. Yeah. And this is what you do. You would know, you teach people about cybersecurity. All right. Cybersecurity analyst, Eric Cole. Thank you so much. Thank you.