00:00:00
Speaker 1: Our tech feature for today.
00:00:01
Speaker 2: We talk about Cybersecurity Awareness Month and how it is that we navigate highly sophisticated AI clones and deepfakes. We're joined by Johan Steen, Human-Centered AI Specialist, 011-883-0702, the WhatsApp line 072-702-1702.
00:00:15
Speaker 1: Johan, welcome to the show. Thank you for joining us.
00:00:21
Speaker 3: Good to be with you, Lebo. Listen, this advert is sponsored by an AI specialist. And then you should also say, can you trust my voice? I've been listening in for a few minutes now.
00:00:32
Speaker 2: Well, I have to ask, how do I check that you are not in fact AI, Johan? What would I need to ask you to check that you're not a bot or a deepfake, that I'm speaking to a human being?
00:00:44
Speaker 3: If I give you a lot of nonsense, which we call hallucinations, then you know I'm an AI. But I'm not. I'm just a dude. So you are safe for now. And it's great to join you.
00:00:54
Speaker 2: Now I'm starting to doubt because I feel like an AI would say, no, I'm just a dude.
00:00:59
Speaker 1: You are safe.
00:01:00
Speaker 3: Yeah.
00:01:01
Speaker 2: So please bring us up to speed with where we are with AI, particularly with those that maybe don't, you know, watch as much content as others. I remember, I don't know if you've seen it, I'm sure you've seen that video that shows the, you know, the transition of deepfake using Will Smith, a video of Will Smith eating spaghetti. And then it goes all the way, I think it was from 2010, it goes all the way to where we are now in 2026. And it looks like, It looks real. You can't even tell if you're not truly paying attention.
00:01:36
Speaker 3: So Lebo, that first video of Will Smith was clearly a deep fake. The shade on the nose was wrong. It looked really bad. But these days, if you see a video on YouTube or on your social media, it is so powerful and so real that it's almost impossible to know whether it was AI created. And This impacts next week, our 4th of November elections and the elections across the world. And people fall for these things. So don't check the sources. But this is manipulation at a massive scale. So don't trust anything you see. Double check it and use common sense because the AI is so good. You will not have an idea whether it's a deepfake or not, sadly.
00:02:21
Speaker 2: Maybe help us understand, what is the intention of deepfakes? Because maybe one can think to themselves, well, if this thing is so harmful, why are we even doing it?
00:02:36
Speaker 3: I think, Lebo, it's all about manipulation. So it can either be a state-sponsored attack from Russia, China, or who knows from who. I don't know if Mauritius will attack us. I'd love to go there again. But it can also be three young people in the garage somewhere because you get free tools. I mean, I've created some deep fakes. Sometimes when I train boards of big banks and other companies, what I would do to start my training, I will create a deep fake of the CEO saying something totally stupid. And they cannot believe it. And I do it to kind of get that punch in of this stuff is happening. I can do it. I do it on free tools. It's so easy to do it. So, I think it's manipulations, either to rig elections for you to buy certain things, to change opinions. Anyone can do it. That's a scary thing. Years ago, when that Will Smith thing came out, you needed a lot of money and expertise to do it. These days, you open a free app, you put in a few prompts, you take a picture of somebody, and you can create anything you want them to say and do. And that is scary, especially for parents, I think.
00:03:47
Speaker 2: Oh, my goodness, and I'm glad you're touching on parents. Before we get to the part about adults being scammed, can we quickly talk about the dangers and what are some of the extreme things people are using with deep fakes and industries, you know, where there's human trafficking and child pornography?
00:04:05
Speaker 3: Oh, there are so many ways, like enticing you to join a new airline as an air host or hostess. I've seen many of that happen, and then you finally fly to Australia whichever country, and suddenly you are being trafficked. But I always say, follow the money. Let me just quickly move. The other phone is ringing.
00:04:26
Speaker 1: They're calling you to tell you you're on the radio.
00:04:28
Speaker 2: Somebody say, I can hear you on the radio.
00:04:32
Speaker 3: And we are coming for you. But it's also, I always say, follow the money. Like, who's making money out of it? We've had some deep fakes on YouTube in particular about an SABC personality.
00:04:47
Speaker 1: Oh, Leon Manas.
00:04:49
Speaker 3: Yeah, and there were some others. And it looks so real and it sounds like them. And people lost hundreds of thousands of rands. Now, I don't want to be ugly because if you're desperate, you go for anything. But it's my view and people shouldn't judge me. But if you fall for this nonsense, you should lose your money. If anything that looks too good to be true, it probably is too good to be true. Check other sources. Google it. Search it. Use your AI. But anything that makes quick money is a scam because no one makes quick money. And now with elections coming up here and then in November later on in the U.S. with the midterms, We're going to see a lot of this stuff flowing around. And also, Lebo, remember, the AI knows from your social media and your online presence your biases. So it will give you a message that will trigger you, whether it is hate toward another gender or another ethnicity. You will see something online where you will go, I told you so. It's those people. It's the Muslims. It's the Christians. And that is dangerous. It's really... I almost think this will lead to the end of democracy as we know it over the next generation. Without scaremongering, this is serious stuff.
00:06:08
Speaker 2: I have to say, I second you on that part because... There's the low-level manipulation of somebody trying to, you know, scare me out of a couple of thousands. And then there's the manipulation, as you said, where our democracy is compromised because people who are outside forces can come in and create any narrative. And because they use online, they use algorithms, and we have a generation of boomers and older who... who just forward messages to one another on WhatsApp. Now a whole belief system is created that alters the decisions that people make and the future of the country. So what can we do about it, Johan? What can we do about it when we know and understand that there's a lot at stake when it comes to AI and also deepfakes?
00:06:59
Speaker 3: Look, this problem will just increase. My advice to people is just think about it a bit. Don't just forward stuff, especially if it relates to your biases or your political or other views. Go and check it. So my wife, and I love her, but she will often send me some sort of a special from some sort of an online company. I do a simple Google search, and the first thing it says is this is fake. So just double-check it. Don't just fall for it. It's an individual responsibility. Now, our children, my son is 12. The stuff that he's getting is shocking. That's influencing. So think before you send it. Double check it. And that's, I think, Lebo, sadly, the only control we've got. Our government should impose regulations. We know that the AI... Regulations have been stalled since April. Hopefully we'll have it next year. But whether there's regulations or not, it doesn't take just a little bit of common sense and being streetwise from each individual. Ultimately, that's my advice. I hope it helps people who are listening in today. Indeed.
00:08:11
Speaker 2: And to always, like you said, Johan, ask yourself why, why now? And who stands to benefit if this becomes a narrative?
00:08:21
Speaker 1: Yo, scary stuff, Johan Stijn.
00:08:23
Speaker 2: Thank you so, so much for your time here on 702 Afternoons and our tech feature. For any of your other tech issues, 702 Afternoons at 702.co.za.
Speaker 1: Our tech feature for today.
00:00:01
Speaker 2: We talk about Cybersecurity Awareness Month and how it is that we navigate highly sophisticated AI clones and deepfakes. We're joined by Johan Steen, Human-Centered AI Specialist, 011-883-0702, the WhatsApp line 072-702-1702.
00:00:15
Speaker 1: Johan, welcome to the show. Thank you for joining us.
00:00:21
Speaker 3: Good to be with you, Lebo. Listen, this advert is sponsored by an AI specialist. And then you should also say, can you trust my voice? I've been listening in for a few minutes now.
00:00:32
Speaker 2: Well, I have to ask, how do I check that you are not in fact AI, Johan? What would I need to ask you to check that you're not a bot or a deepfake, that I'm speaking to a human being?
00:00:44
Speaker 3: If I give you a lot of nonsense, which we call hallucinations, then you know I'm an AI. But I'm not. I'm just a dude. So you are safe for now. And it's great to join you.
00:00:54
Speaker 2: Now I'm starting to doubt because I feel like an AI would say, no, I'm just a dude.
00:00:59
Speaker 1: You are safe.
00:01:00
Speaker 3: Yeah.
00:01:01
Speaker 2: So please bring us up to speed with where we are with AI, particularly with those that maybe don't, you know, watch as much content as others. I remember, I don't know if you've seen it, I'm sure you've seen that video that shows the, you know, the transition of deepfake using Will Smith, a video of Will Smith eating spaghetti. And then it goes all the way, I think it was from 2010, it goes all the way to where we are now in 2026. And it looks like, It looks real. You can't even tell if you're not truly paying attention.
00:01:36
Speaker 3: So Lebo, that first video of Will Smith was clearly a deep fake. The shade on the nose was wrong. It looked really bad. But these days, if you see a video on YouTube or on your social media, it is so powerful and so real that it's almost impossible to know whether it was AI created. And This impacts next week, our 4th of November elections and the elections across the world. And people fall for these things. So don't check the sources. But this is manipulation at a massive scale. So don't trust anything you see. Double check it and use common sense because the AI is so good. You will not have an idea whether it's a deepfake or not, sadly.
00:02:21
Speaker 2: Maybe help us understand, what is the intention of deepfakes? Because maybe one can think to themselves, well, if this thing is so harmful, why are we even doing it?
00:02:36
Speaker 3: I think, Lebo, it's all about manipulation. So it can either be a state-sponsored attack from Russia, China, or who knows from who. I don't know if Mauritius will attack us. I'd love to go there again. But it can also be three young people in the garage somewhere because you get free tools. I mean, I've created some deep fakes. Sometimes when I train boards of big banks and other companies, what I would do to start my training, I will create a deep fake of the CEO saying something totally stupid. And they cannot believe it. And I do it to kind of get that punch in of this stuff is happening. I can do it. I do it on free tools. It's so easy to do it. So, I think it's manipulations, either to rig elections for you to buy certain things, to change opinions. Anyone can do it. That's a scary thing. Years ago, when that Will Smith thing came out, you needed a lot of money and expertise to do it. These days, you open a free app, you put in a few prompts, you take a picture of somebody, and you can create anything you want them to say and do. And that is scary, especially for parents, I think.
00:03:47
Speaker 2: Oh, my goodness, and I'm glad you're touching on parents. Before we get to the part about adults being scammed, can we quickly talk about the dangers and what are some of the extreme things people are using with deep fakes and industries, you know, where there's human trafficking and child pornography?
00:04:05
Speaker 3: Oh, there are so many ways, like enticing you to join a new airline as an air host or hostess. I've seen many of that happen, and then you finally fly to Australia whichever country, and suddenly you are being trafficked. But I always say, follow the money. Let me just quickly move. The other phone is ringing.
00:04:26
Speaker 1: They're calling you to tell you you're on the radio.
00:04:28
Speaker 2: Somebody say, I can hear you on the radio.
00:04:32
Speaker 3: And we are coming for you. But it's also, I always say, follow the money. Like, who's making money out of it? We've had some deep fakes on YouTube in particular about an SABC personality.
00:04:47
Speaker 1: Oh, Leon Manas.
00:04:49
Speaker 3: Yeah, and there were some others. And it looks so real and it sounds like them. And people lost hundreds of thousands of rands. Now, I don't want to be ugly because if you're desperate, you go for anything. But it's my view and people shouldn't judge me. But if you fall for this nonsense, you should lose your money. If anything that looks too good to be true, it probably is too good to be true. Check other sources. Google it. Search it. Use your AI. But anything that makes quick money is a scam because no one makes quick money. And now with elections coming up here and then in November later on in the U.S. with the midterms, We're going to see a lot of this stuff flowing around. And also, Lebo, remember, the AI knows from your social media and your online presence your biases. So it will give you a message that will trigger you, whether it is hate toward another gender or another ethnicity. You will see something online where you will go, I told you so. It's those people. It's the Muslims. It's the Christians. And that is dangerous. It's really... I almost think this will lead to the end of democracy as we know it over the next generation. Without scaremongering, this is serious stuff.
00:06:08
Speaker 2: I have to say, I second you on that part because... There's the low-level manipulation of somebody trying to, you know, scare me out of a couple of thousands. And then there's the manipulation, as you said, where our democracy is compromised because people who are outside forces can come in and create any narrative. And because they use online, they use algorithms, and we have a generation of boomers and older who... who just forward messages to one another on WhatsApp. Now a whole belief system is created that alters the decisions that people make and the future of the country. So what can we do about it, Johan? What can we do about it when we know and understand that there's a lot at stake when it comes to AI and also deepfakes?
00:06:59
Speaker 3: Look, this problem will just increase. My advice to people is just think about it a bit. Don't just forward stuff, especially if it relates to your biases or your political or other views. Go and check it. So my wife, and I love her, but she will often send me some sort of a special from some sort of an online company. I do a simple Google search, and the first thing it says is this is fake. So just double-check it. Don't just fall for it. It's an individual responsibility. Now, our children, my son is 12. The stuff that he's getting is shocking. That's influencing. So think before you send it. Double check it. And that's, I think, Lebo, sadly, the only control we've got. Our government should impose regulations. We know that the AI... Regulations have been stalled since April. Hopefully we'll have it next year. But whether there's regulations or not, it doesn't take just a little bit of common sense and being streetwise from each individual. Ultimately, that's my advice. I hope it helps people who are listening in today. Indeed.
00:08:11
Speaker 2: And to always, like you said, Johan, ask yourself why, why now? And who stands to benefit if this becomes a narrative?
00:08:21
Speaker 1: Yo, scary stuff, Johan Stijn.
00:08:23
Speaker 2: Thank you so, so much for your time here on 702 Afternoons and our tech feature. For any of your other tech issues, 702 Afternoons at 702.co.za.
