Transcript of AE 1390 - The Goss: The Internet is Drowning in Fake AI Video Slop

G'day, you mob! Pete here, and this is another episode of Aussie English, the number one place for anyone and everyone wanting to learn Australian English. So today I have a Goss episode for you where I sit down with my old man, my father, Ian Smissen. And we talk about the week's news, whether locally Down Under here in Australia or non locally overseas in other parts of the world, okay. And we sometimes also talk about whatever comes to mind, right. If we can think of something interesting to share with you guys related to us or Australia, we also talk about that in The Goss.

So these episodes are specifically designed to try and give you content About many different topics where we're obviously speaking in English and there are multiple people having a natural and spontaneous conversation in English. So it is particularly good to improve your listening skills.

In order to complement that, though, I really recommend that you join the podcast membership or the Academy membership at AussieEnglish.com.au, where you will get access to the full transcripts of these episodes, the PDFs, the downloads, and you can also use the online PDF reader to read and listen at the same time. Okay, so if you really, really want to improve your listening skills fast, get the transcript, listen and read at the same time. Keep practising and that is the quickest way to level up your English. Anyway, I've been rabbiting on a bit. I've been talking a bit. Let's just get into this episode guys. Smack the bird and let's get into it.

Dad, what's going on?

Hey, Pete. How do you know what's real?

Uh, that's a deep question.

It is a deep question.

I guess you don't.

You don't? Well, and that appears to be the case.

Yeah.

Now, there was an article recently on the ABC, about. Well, the headline is teenagers struggle to tell if videos are real or fake as AI flood social media.

I think, I don't think it's unique to teenagers.

I don't think it's teenagers. I think the teenagers, well, they were high school students that were asked.

Yeah, if they're finding it hard. I think the rest of us are.

Well, exactly. Because they're the ones usually creating them.

Yeah.

So, yeah. And then, I mean, the story is that's the story. You don't need to, you know, there's no.

There's an experiment there we showed for high school students. Wow. What a big pool of. Yeah, 21 short videos, ten of them real, 11 AI generated. After each of the videos, we asked the students to decide if what they watch was real or AI. Before the test even started, the students were not exactly confident in their ability to spot a fake.

Yeah.

So, yeah. They found it difficult. It's one of those things I feel like I'm not too bad at it, but it really depends on the context. Usually it's pretty the, we're still at that stage where a lot of them are still very obvious. Especially if there's a lot of movement or stuff going on. Like if you've got, I don't know, there's the one video I've seen of like the gorilla charging at the, at the kids and smashing the glass.

Yeah.

And it's, part of it is like, I know gorillas don't really just do that. Behaviourally. The other part is like the amount of damage done to the glass.

Yeah.

But they. A lot of them do look very convincing.

Yeah, a lot of them. A lot of the ones where you've got that sort of thing can look convincing. The ones that don't look terribly convincing to me are the ones which are like everything from sort of full body shot down to head shot of people because they're very stylised.

Yeah.

And you know, the, the number of random sort of attractive young women stuff that appears in of just spam on, on social media.

Yeah.

And you scan through it and you go, Fake. Fake. Fake. Because they all look the same!

Part of me is so surprised at how many there are now being done by like, pages on Facebook. I try to think, what is the monetary.

Yeah,

Incentive here because people will just make fake photos and then have a story about like and you know, there was one that I saw that...

Click through on Facebook doesn't give you anything.

Yeah. There was one I saw today that was like lion had caused local extinction of meerkats. And it was very clear that the pile of dead meerkats was fake.

Yeah.

In the image. And so I'm like. And the story, what was the story? The the line had the the tusk of an elephant stuck in its leg, and when a local vet removed it in order to reward the vet, you know the line went and killed all these meerkats. And I'm like, well, that's just a bullshit story.

Yeah, they...

Just would never happen. There are so many levels to that. How would a tusk from an elephant get stuck in a lion's leg? And why would the lion just go and massacre random meerkats and not eat them?

I know!

But then, yeah, they accompany it with a video or something, and you sort of like, who is this for?

Yeah.

Is it just people wanting to be incredulous and, like, believe all this stuff and. Credulous!

Yeah, I understand people, you know, don't obviously don't condone it, but I understand why what the motivation is for people to create fake accounts on these things, because usually they are trying to prep the prep their audience for a scam.

Or just mask their own identity, right.

Yeah. But, but but often it's prepping people for a scam. It's like the how many celebrity fake celebrity things pop up all the time. Hi. You know, I'm Chris Hemsworth, you know?

Yeah. I gotcha.

Yeah, I want to, you know, I'm following you. Yeah. Chris Hemsworth never following me on social media.

Let alone sending you a video to let you know.

I know, I know, and you and you sit there.

Hey, Ian!

Really? There's like. But I understand why people are doing that because the next thing that comes from them will be a post that broadcast to the 10,000 people who friended them that says, hey, I'm raising money for blah, blah, blah. You know, would you, you know, send it to this bank account?

Is this something that's sort of freaking you out, though? Like, as much as it is kind of still, you see a lot of these videos and you're like, this is just slop like, and I'm not going to be fooled, you know, blah, blah, blah. Does it get to the point where you're starting to like, be like in, in a year? Where is this technology going to be in ten years? Where is it going to be?

Yeah, there's the technology doesn't particularly worry me. What worries me is that fortunately, most people are cynical, and so that means that the people who are not are much more likely to be scammed by things. But also, we will just not trust anything.

That's what I keep...

People who are cynical are going, I don't trust anything.

Yeah, anytime I'm on Facebook, Instagram now and I see those sorts of stories.

I immediately Google it.

Yeah, exactly.

This never happened. This never happened. This never happened. Yeah. It's and a lot of them are, fake advertising for things.

Yeah.

Like they'll, they'll use the, you know, such and such from morning television recommends you do blah, blah blah. And you Google they never did. But it's they've AI-d up, you know, somebody's looking like they're and they never actually have them talking directly about it, but they'll mock it up. It looks like this was a news story from some time that you just missed on the channel whatever news, you know.

What do you, what do you think it means, though, for things like. What would you say, trials in the future, or evidence? Because I imagine now with AI like ChatGPT and everything, you probably just, you could manufacture video files that one have the video showing something that didn't really happen and how that could impact, you know, a court case, but it could manufacture the data to pretend like it was authentic, right. You could say, make this file appear as if it was filmed in this location at this time, uh, you know, and so that it can appear like evidence for this thing that did happen or didn't happen, right. You know, so I wonder how much that'll impact things to where if you previously, if you had court cases, people could bring video evidence and be like, here's this person doing this thing. And based on the metadata in this file.

Or the reverse, here's this person, not in another state at that time. And if you can manufacture a picture of me in a bar in outback New South Wales. Date stamped a particular thing.

Yeah. It's like, what happens when, when this, this technology effectively makes something like video evidence mute and useless and pointless. You know, it's a weird, it's weird that they could suddenly take away a big part of what is used in criminal cases...

Or what we have accepted...

Or news where you, you know, you would you trust your eyes. But in the near future you will probably be so sceptical.

And a lot of it too is. And this is, you know, one of my big bugbears about people believing stories without evidence. And the old gag about, you know, anecdotes are not data.

Was it the plural of anecdote isn't data?

Yeah.

Unless you're studying anecdotes.

Yes. Exactly. But you know where and even. But then, if the anecdote is fake, that makes it twice as bad. But people are. But you know my mother's, you know, friend's uncle said this. Well, it must be true.

And.

I choose to believe what I was programmed to believe. There's a Futurama quote. Yeah. It's going to be an interesting world. I keep thinking about just the impact that this technology is going to have on, on jobs, on critical thinking, logic, all of these things for our kids in particular, because it's I, I'm coming to realise more and more and more, my generation is the last one to have kind of lived prior to the internet and then to have passed through that period of all of this technology kind of, you know, being released to the public, the internet, computers, smartphones, all these things and getting used to it.

Yeah.

You wonder if kids these days are going to be, if they'll have that same sceptical, sceptical nature or you know.

Well hopefully, hopefully the, this actually increases it.

Yeah.

The critical thinking part is going to be the challenge. And, and you know, I've always I mean, I used to be a high school teacher, and I always used to say to kids when they come up with stories, oh, look, this was in the news. And why would you. What do you mean by the news? Why would you believe this?

Yeah.

Just because it's being reported on the news, whether it be in the news, you know, the TV media, radio, newspaper or whatever, what's what's the actual evidence, what's the reliability of the source that you're getting it from and the source that they are getting it from? What's the motivation for this? And the motivation with news media now is clicks. It's advertising. It used to be the motivation for old fashioned news media. Used to be keep you watching the half hour television show or buy the newspaper because the advertising is paying for that to be done with online media. These days, it's just click through, click through, click through. If I'm looking at this for two minutes and scrolling through, the advertisers know that I've been on there for two minutes and they will pay more to that news source...

It's, the sad thing...

Have it on there and then...

It's incentivising poor quality sensationalist AI slop news, too. So you'll read it quite often and be like, this is just such.

It's just rubbish.

It's just trash that's been, it's been designed to rage bait me into checking it out.

Exactly.

And then it turns. It's so often now too. My habit is getting on Facebook or Reddit or whatever, and the link will say a certain thing and go to a news story. And then I'll go to the first comment and be like, because it's always. I'll save you the click. Here's what the story actually is. And you're like, that's enough for me.

And that's the point is the headlines now are all rage bait or clickbait. They're all seducing you to go to the thing because as soon as you click on it, then the advertisers are more likely to pay money to that thing.

And this is the biggest hate I have for my job, is that I have this constant battle with YouTube or with podcasts. Between creating sensationalist clickbait titles that are designed to get as many people to click as possible and reflecting the reality of whatever the content is, and you have to draw. Obviously it's a spectrum between complete rage bait, clickbait shit, that is not true, but we'll get people, you know, oh, Princess Diana is still alive.

Yeah...

Elvis...

Elvis lives!

Whatever it is, you know? Complete bullshit. And then the video is nothing about that. To, to having a very, very boring and bland literal title for the video that everyone's going to look at and be like...

Yeah, why would I click this?

And it's hard finding that balance of like, you have to clickbait enough that people generate that curiosity to want to check it out, but not so much that when they watch it, they feel like you didn't deliver,

Or...

And they get frustrated..

They watch for 15 seconds and go gone.

Yeah, exactly.

You know.

Yeah. Well, and that's the tough, the tough thing in this landscape of online media content is trying to.

Get them there and hold them.

And be ethical about it.

Yeah.

That's the hardest part. And it can be a race to the bottom, which I think has happened for a lot of media companies because they rely solely on, on generating ad revenue now.

Yeah.

It's like we just need as many people clicking on and seeing the ad for this amount of time to get paid for it. And they don't. It doesn't matter if they stay or not. It doesn't matter how accurate the stuff is.

The, the irony is that the alternative to that is a subscription service to news, which a lot the Murdoch media in Australia. You cannot read news articles on Murdoch media without subscribing. And you go, all right, well, that theoretically means.

I'm not paying to be brainwashed!

Exactly! But that theoretically means that they, they're relying on the quality of their stories for you to stay subscribed.

Yeah.

But every single story you see is clickbait. And you go, really? Why would I pay to watch this rubbish and, or read this?

Do you think it'll get to a point, though, where like, as you were saying, maybe it'll, it'll do the opposite of what I was worried about. And kids will just become skeptical of everything. Super skeptical. But I guess it can become. It's sort of almost a nihilist side of it, where at some point everyone's like, well, what can nothing. It's all..

Nothing is real.

Nothing is true.

Yeah.

I guess it's that post-modern sort of thing. Why even try? Which is interesting because I guess that's kind of like what happens in some foreign countries, where the government controls the media.

You just accept whatever you're given or not accept it because there's nothing you can do about it.

Quite often they will intentionally put out there both sides of the story or both extremes, in order to just exhaust the public so that they just tap out and say, well, I'm not going to bother. I'm not even going to bother trying to work out what's true and what's not. So I find it interesting thinking about that. Like, our kids are going to get to the point where they become super skeptical and logical and constantly evidence based and like, I want to get to the bottom of it. Or do they go the other way where they're just like, I don't give a fuck and it's too hard.

Just give up?

Yeah, you just be like...

Well, hopefully. Hopefully. Yeah. And look, I think the answer from an adult grandparent in our point of view is, is encouraging. I hesitate to say teaching, but encouraging kids, to be critical.

And curious.

And curious. And wanting to work out and go, well. And I said it's evidence and motivation. What evidence are they providing? What's the sources? How reliable, how reliable are they and what's the motivation for them? Um, and if you start to read and examine everything that you see and are experiencing that way, um, and even even to the point of when people are telling you things, they go, well, what makes you think that's true?

I know so often I'll be sharing posts and then before I click share, I'll be like, is this just something that confirms my bias?

Yeah.

Or is it bullshit? And I need to actually make sure through other sources this is true because so often you'll read something and be like, oh yeah, that that has to be true. I want that to be true. You know, that agrees with what I think you know, and then you're like, ah, but is, is it? Like the lazy thing to do would just be to take it on face value and share it and, and now and partly, too, I'll just be like, well, I can't be getting into it, so I'll just not share it.

Or I'll just share it and have people tell me I'm an idiot. Yeah.

Yeah, that's it. And then just apologise. Oh, well, I didn't realise it.

Yeah.

All right. Cool. So there you go. Yeah, it's probably only going to get worse. The interesting thing too is thinking about, as a final question, do you think that individuals have a right to their appearance and that there should be laws that prevent AI being able to use other people's likenesses? I guess, in videos?

Yes.

Do you think that will come through? Is that just a matter of time, before those laws...

Well, the irony is that, the irony is that there are laws in many countries Australia...

Yeah.

US, UK, don't know about many European countries. I suspect there are, they're the same. There are laws about that regarding.

Like impersonating people.

No, but ignoring the AI thing, like, my image. You know, you cannot use my face on, for commercial purposes without my permission.

Yeah.

You can't use my voice without my permission. Now is an AI generated image of me or an AI generated voice of me reading or telling something. Is that me? And that's where, that'll, that'll be tested in court. And it won't be long before some celebrity goes. That's not me.

Yeah.

They're representing it as me..

It'll be some big company. Will, will decide to use an AI video of someone famous.

It'll be. And I'll use the Chris Hemsworth thing before. And Chris Hemsworth just going to go. I can sit there. I can put tens of millions of dollars behind a legal thing. And to just do.

Like, hundreds of millions..

You're screwed.

Yeah.

Yeah.

Exactly. It'll be interesting to see when and if that happens.

And it will be somebody. Taylor Swift. It'll. Yeah.

Well, the rest of us don't have the money to do that.

Exactly. And, and I mean, Taylor Swift is a good example of somebody who, who ended up, she took somebody to court for sexual harassment in a job when she was very young.

Yeah.

And they countersued her for defamation, and she won a dollar.

Yeah, she didn't actually won..

She didn't care. She didn't want the money. She just wanted to make the point that you're going to spend a lot of money going after me to defend yourself for doing something that was wrong.

Yeah.

And, and I think that's going to be one of the, and that's going to be the precedent that sets these things apart. Where you cannot artificially generate what appears to be somebody's voice or image.

Yeah. For, for financial gain.

Yeah.

Yeah. It is interesting that the apps like Sora and these other ones that generate these videos at the moment can just use intellectual property of other companies and people.

Oh yeah.

Ubiquitously. Like I think that initially when, and they probably did it intentionally, when they released it. You could make videos of all these Nintendo games and characters and other things, and then all of a sudden Nintendo was like, ah, excuse me. And they disabled the ability to do that all with Pokemon or something.

Yeah.

And you're sort of like, you're like, okay, so you can do it for that, but there's all these other areas that you're not doing and it's like, yeah, it's just gonna have to be whether or not you have the money to follow it up, right?

Yeah.

Anyway, yeah. We're fucked.

See you guys!

Loosely. Bye!