Angry Planet - Navigating Reality and War During the Age of AI Propaganda

Episode Date: June 5, 2026

On the morning of April 21, Trump posted an image of eight women on Truth Social, claimed they were Iranian dissidents set to be executed, and demanded that Tehran release them. Detractors, and several Iranian sources, claimed the women were AI-generated. A day later Trump claimed the women would no longer be executed and that he’d saved them.The truth is that the women are real and many are still in danger. Trump’s post made real Iranian women who protested the Iranian regime appear fake. The story speaks to a moment we’re in where it’s become impossible to parse truth from lies online. This was already difficult before AI-generated pictures and video. Now it feels impossible.On this episode of Angry Planet, Mahsa Alimardani is here to tell us the story. Alimardani is the Associate Director of Technology Threats and Opportunities at WITNESS.Eight real women turned into AI propagandaReal crimes bastardized into regime propaganda“We need to come to terms with the fact that our information environment is structurally different.”Content Credentials as a partial solutionHow AI is supercharging our chosen reality tunnelsThe cycle of uprising and repression in IranThe structure of Iran’s internet and how its blackouts workDomestic intranet as an alternative form of communicationAI-generated Lego propaganda videosIran ReframedExplosive Media’s deep connections to the Islamic RepublicPolitics as fandom, fandom as politics“Everything is becoming flattened.”“The onus on the person scrolling is a bit unfair.”Mahsa Alimardani’s LinkedInThe Real Iranian Women Protesters Trump Made Look SyntheticIn the Room With Iran’s Social Media SavantsHow AI Content Detection is Being Weaponized in the Iran WarIran Is Winning the AI Slop Propaganda WarSupport this show http://supporter.acast.com/warcollege. Hosted on Acast. See acast.com/privacy for more information.

Love this podcast. Support this show through the ACAST supporter feature. It's up to you how much you give, and there's no regular commitment. Just click the link in the show description to support now. Hello, and welcome to another conversation about conflict on an angry planet. I am Matthew Galt. We are back. Yes, there are more shows happening.

Lori not. I'm still alive. We're going to talk about something today that has been kind of driving me a little crazy. I think it's been making a lot of people in my life crazy, and I can't think of a better guest to kind of get into all of it. She's been covering it for quite some time and quite well. Will you introduce yourself? Yeah, hi, Matthew.

Thanks for having me. I'm Masa Ali Mardani. I am the Associate Director for Technology Threats and Opportunities at the Human Rights Organization Witness, which is preoccupied with defending reality, audio, visual evidence in the age of AI. And I have been kind of working in the digital rights realm for, I guess, almost 15 years now and doing academic and soul society work for a while. And I am Iranian and I have studied, you know, the Iranian internet and Iranian communications for a while as well on top of all that other stuff that I do.

preoccupied with defending reality is quite the turn of phrase. Yeah, yeah, maybe over promising our mission. But yeah, we try, especially in the realm of audio-visual content, which has been what witness has done, you know, even before the age of AI. It's been about using video for accountability and human rights documentation. Somebody has to do it. And one of the reasons I wanted to have you on is I had this interaction last month where a friend sent me a tweet from or a truth social post from Donald Trump. And, you know, say him screencapping somebody else's account on X.

There's eight photographs of women kind of looking direct at camera. And these are, we're Iranian women. that Trump said were, or the original post said and Trump was then broadcasting were set to be executed by the Iranian regime. My friend shows me this and he says, can you believe that the president is passing off these AI generated photos is real? And I looked at it and I said to my friend, like, are we sure that that's what's actually going on here? And I started to dig around a little bit and discovered that they were very real women, but also that that is, just like a part of the story. And I kind of got, and I was having a busy day,

and I got to this point kind of looking into this, where I think a lot of us get, and you've written about where I just tossed up my hands. And I backed away, and I said, I don't have the bandwidth right now to figure out what's truth, what's fiction, what's going on. And that is what you do. That is where the most of your bandwidth goes to

is sussing out truth from lies and also kind of analyzing what AI is doing to us right now and what digital technology is doing to us right now and why so many of us are throwing up our hands. So, I mean, we can back all the way up to like April 21st and the original post on X. Can you kind of walk us through the beats of what happened over like that 48 hours? Do we start on the day of the post or do we start at, I don't know, there's so many places you could start with coming to this story, right? I think for our audience, let's start on the day of the post.

Because I think this is a, I'm going to say a really messed up sentence. But I think that initial post on the 21st and then Trump, broadcasting it on truth social, is maybe when some of this, people start paying attention to some of these things, if that makes sense. Right. Yeah. So, I mean, that day, when he does post,

it's in the middle. of the ceasefire negotiations that are going on between the U.S. and Iran. And throughout this whole conflict, there has been this narrative, you know, both from Israel and the U.S. that, you know, the human rights of Iranians, the freedom of Iranians is part of the story. It's part of why the conflict is happening. And, I mean, I think we all know, given the human rights situation in both these respected countries that human rights isn't often central and it's a little bit insincere to consider that they would care about the human rights of Iranians or Iranian protesters. But this has been very central to the narrative.

It's been central to a lot of the war propaganda, of course. And, you know, we can unpack the war propaganda coming from the side of the Iranian regime as well. But this has been part of the story. and what is really frightening. And, you know, like, the reason why I work in the human rights field, why I do work on tech and human rights is because I am Iranian and I have spent a lot of time being personally impacted by, you know, what has happened in Iran at the hands of this regime.

And I know firsthand how, you know, the atrocities the regime has been responsible for. So in this moment to see, you know, this topic that I have worked on for many years, that I know many people work hard to, you know, authentically document and provide truth and accountability about the crimes of the regime to see it kind of become bastardized by the U.S. government, by Trump. It's really devastating. And this kind of layer of what AI is doing to the story is really deteriorating this field. And I mean, just from what has happened from the start of the war till now in terms of just the ability to do human rights documentation, to do any form of justice and accountability of work on the side of the victims of the Iranian regime has been so terribly undermined by what the U.S. and Israel have done.

and watching that post in real time garner that outrage, garner the disbelief that this regime could be doing terrible things, could be arresting protesters. Because, of course, you know, the story of that post was that he was trying to leverage the negotiations to stop the executions of these eight women. And these eight women were real. the photo he shared were AI-enhanced AI-edited photos. But the information he was sharing was incorrect. One of eight of those women has been convicted of a crime that can lead to execution.

So she is up for execution. It's not an imminent execution yet. The other seven women are not up for an execution center. However, a few of them are still in jail, a few of them are out on bail, but they were all, you know, arbitrarily arrest for enacting their right to protests throughout the January protests. And so what this has done has basically undermined our efforts to document what this regime has been doing to protesters, the fact that there are real protesters facing execution. there have been several of the protesters from January, in fact, executed in this interim period, but there is now just a general disbelief that these crimes are happening at the hands of the Iranian regime because the president of the United States shared not only AI-enhanced images,

but disinformation about real women. And, I mean, we've been following this kind of accusation that, you know, it's all AI-slop. for a while. And, you know, this phenomenon of the liar's dividend of, you know, just general disbelief, especially being harnessed by bad actors to, you know, hide inconvenient truths by using the accusation that everything is AI. So this has been really just in the atmosphere and in the environment, and it's been going into overdrive during this conflict, just this sense of uncertainty. about what is real and what is not real. And I do really believe this conflict has led to, you know, victims of the Iranian regime just being severely undermined because of the way that this conflict has gone forward with the narratives that the U.S. and Israel have created. And Iran has leaned into this in a very effective in the West way, right?

they, multiple regime-connected accounts kind of said, oh, he's sharing AI slop. And what did the judiciary put out said that none of these women were, had been arrested, right? What exactly did they say? Yes, the judiciary, I mean, the Iranian judiciary has a media arm that is very active and immediately said that this was fake news. These women didn't exist when in fact they do exist. And we have actually seen a lot of these. back and force between a piece of AI content going wild and the Iranian judiciary reacting to the piece of AI content to basically say, oh, these are all lies, this isn't actually

happening. It's happened at the very start of the war with another case. I was following when, I mean, during the June 2025 war, when Israel started the war, we saw genuine concern for the lives of political prisoners, especially since a lot of the prisons were in places that were getting bombed. And in fact, one of the most famous prisons, Evan Prison, was bombed by Israel. And real prisoners became casualties and victims. And it was called a war crime by a human rights watch. And under Geneva Conventions and international law, political prisoners need to be freed or put into safe locations during wars. and of course the Iranian regime does not care about the safety of political prisoners and did no such thing.

And at the start of the war again, human rights organizations were sounding the alarm on the safety of political prisoners. And somehow an AI video started getting circulated of political prisoners being moved and used this human shield at a Sepah and IRC military base. in Tehran. I remember someone sent me this video and they're like, Masa, what's up with this video? I, like, looked at it quickly. I thought it looked real because I was out. And then I sat up my computer and listened to the audio and it was very clearly like AI Persian. Like AI Persian audio is very obvious. And I noticed that a very well-known Iranian media in the diaspora with millions of followers was sharing this AI video as if it was real to peddle. the kind of narrative that political prisoners

not only were endangered in the prisons, but were being used as human shields. And the Iranian judiciary immediately reacted to this piece of AI of like, see, they're trying to make us look like we're endangering the lives of political prisoners. Political prisoners are perfectly fine. All the prisoners are fine.

While, like, it was being very actively documented what the real dangers were. And so this back and forth has existed for a while, And of course, the Trump incident is the best known one, obviously, because it made world headlines immediately. Another interesting aspect of this, something that you've written about before, is kind of this weaponization of AI detection tools. And in this instance, you have these eight photographs of these women, and their faces have been doctored. Maybe he's not quite the right word, but touched up in a way that basically every photo of anyone is touched up. now. A lot of it kind of runs automatically when you take pictures with your phone.

And if you run it into some of these, like you just pick a website and toss it in and says, is this AI, it will ping on an AI detection tool. And it's not necessarily that the photograph is completely fake. It's just that it was, it had a layer of AI gloss added to it. And that's something we've seen happen also. So, you know, photos get run through, run through these tools. And then the tools themselves are kind of used to say things are fake, right? Exactly, right.

That's exactly right. And to be fair, the AI detection is doing its job in this instance. Because there are AI artifacts. AI is being used. It's just that we need to come to terms with the fact that our information environment is now structurally different, which is AI is available in a lot of processes that people would use, right? were used to photojournalists editing their photos for lighting or contrast or different things. We're used to people designing and sharing photos with some element of editing.

And now the reality is a lot of editing software has AI in it. And I mean, there's been so many different cases where we've seen this. During the protests in January, there was a famous kind of symbolic photo of a protester that went viral. And it was blurry. so someone used AI enhancement to make it clear. Immediately, people lashed on because you could see some AI artifacts that this is all AI slop. Protests aren't actually happening. And of course, this was a regime narrative that Iran used.

And I think there's just a general misunderstanding that this is our kind of AI reality. When it comes to the Trump images of the AI enhancement, it is very dangerous because basically it's feeding into this general AI doubt that we have, this general misunderstanding that this is the capabilities that we have. I mean, we have this kind of service within my team at witness. It's called the Deep Fake Rapid Response Force, where you have different people escalate, audio, images, videos to see if there's any kind of AI involved. And it's often very hard to be able to say that this is 100% AI or this is partially

enhanced with AI. And the thing that we do actively advocate for within my team is to create trust signals with things like content credentials and content credentials give you provenance, which is kind of the recipe of how a piece of content is made. Of course, it's very difficult to restructure our entire information environment, but it is something that we increasingly push for all of the platforms, for the kind of models behind content creation, behind the editing software. Actually, Adobe is one of the biggest kind of tech companies behind content credentials,

and they do offer content credentials across their entire editing software. And so, yes, content credentials is part of this kind of coalition of different tech companies and capture devices that can enable these kind of cryptographically enshrined provenance metadata that can give you the recipe of how a piece of content is made. Obviously, it's going to take a while for that to be enforced throughout the information. environment, but it is something that I think does have a potential to act as one potential trust signal in our increasingly, you know, uncertain world filled with disinformation. What do you think that that would look like from the user's end?

You know, if I'm on, say, X and I'm doom scrolling and I'm just cycling through and I'm looking at photos trying to parse like, okay, is this AI generated? Is it real? Who's this poster? you know, things are happening very quickly. I'm having emotional reactions to the images I'm seeing. Like, what does, what do, like, how do the credentials help in that situation? So in the ideal world, which I know we might be far from coming to, in the ideal world, for example, you saw the Trump photo.

You would see that, okay, these are eight different photos that originated from, like, these kind of eight different originals captured on these devices and now they've been put through they've been first of all collated into one collage and gone through like x and y editing or this filter process and it would give you that story of how that has come to be uh you you do see some of these examples now um if you have a google pixel 10 you can basically see it see the content immediately that it was taken with the Google Pixel 10 camera. Obviously, it will take a long time for, you know, all the phone companies and capture devices to enable something like this. But there is potential. I mean, in like one example that we did keep seeing coming up was actually from photo

journalists and from wire services during this war. And so the interesting example is that, of course, in Iran, journalism, media is very centralized and controlled by the authorities. And when you do have a case of protest massacres where victims are being created at the hands of the authorities themselves, they go out of their way to stop documentation. They don't want any witnesses to those events. When we saw the war happening, when we do see kind of victims' civilian casualty at the hands of U.S. and Israeli bombs. The authorities are invested in documenting this, right? They make sure the photojournalists go to those sites and take photos.

And so, and I always say this, you have to hold two truths at the same time, which is you do have civilian casualties, and you do have authorities who are over-invested in documenting it and potentially mobilizing that documentation for propaganda. It doesn't mean that that's not real, that we don't have civilian casualties. And so what we saw was these photos captured by kind of regime or state-affiliated photo journalists sold to international wire art services, published in The Guardian and the

independent, etc. And then we had people accusing those photos of being AI generated. And this was going viral to the point that I think there was a period of like five days at the beginning of the war that I had, people in my personal network sending me these photos, being like, what are we supposed to believe? Are civilians actually dying? This looks like it's all AI,

and it was people legitimately referencing some sort of, like, forensic-looking analysis to say that these photos, these real photos that were verified were AI. Now, in that ecosystem, if the photojournalists, if the wire services required the photojournalist to enable content credentials, which very easily could be enabled right now, because most photo journalists are doing touch-ups, are using some sort of editing software to fix their photos

when they sell them to wire services. And right now, you can enable content credentials through that editing software. And if wire services made that a kind of journalistic standard, when this photo went out, you could potentially turn to a content credential when it is posted online to see like, okay, this was captured on the X device. And then it went through this like editing software to get to this place that it is with the like color contrast changed.

I'm imagining because like the, I'm imagining what that would even look like if I like at the end point, we have a true, a Trump social post. that is a screenshot of an ex post that is a collage of eight different AI touched up images that were taken. Who knows where. And I imagine like being able to click that truth social post and then getting like this diagnostic breakdown of like each individual piece and kind of the chain that it all

went through without me having to do it, which is exhausting. So yeah, I think that that would help. I worry that people wouldn't avail themselves of it, but I think at a certain point, like, you can lead a horse to water, right? And you can't quite control for what everyone is going to do in the moment. And I think that's another part of this is that I've been very struck by any of this has been true for as long as humans have been around. But people tend to just believe the things they want to believe. They kind of have this narrative, like, in their head. And they kind of pick the evidence that suits that.

And I think that both or the U.S. and Israel and Iran all know that. And they're tailoring their propaganda, all three of them, in a way that makes that easier for people to do, right? Yeah, 100%. And, I mean, I talk about this conflict of being really the worst case scenario of, like, all the things we predicted AI was going to deteriorate in the information space. I don't think we quite had arrived at it, but I do think Iran is this laboratory, right, where you have so many different stakeholders with their positions. Like, as someone who is in the Iranian diaspora, I see it firsthand, right?

I'm part of diaspora that has generally been, and I'm not saying every Iranian is, like this, but we're very critical. I am part of a community that's very critical and has seen the harms of the regime. And so there is this natural tendency to not believe the regime. You know, there's a massive part of the Iranian opposition that has a lot of sympathy towards Israel, because of the, you know, the dynamics of the enemy of my enemy. And so for this community, Israeli propaganda, Israeli disinformation and influence operations is very, this is the perfect and ripe opportunity to do influence operations as Israel. And they have been quite effective. There's been a lot of kind of research that's documented this.

You can see a really good citizen lab research from last summer that's kind of documented some of these influence operations through the Iranian opposition. And then, of course, you have, like, the Iranian regime. They obviously know what, like, anti-war sentiment is like. They know what leftist movements, how sympathetic they are to, you know, the Palestinian cause, to anti-imperialist criticisms of U.S. and Israel. They also know this very well, and they know how to influence from their side. And so there's just so many different opportunities to muddy the... the water because there's so many different positions and so many people willing to want to

believe in the things that they already believe in. Iranian opposition is already willing to believe that the Iranian regime are liars, that they hide things. And so, like, this disbelief in civilian casualties. And to be honest, I see it to this day amongst some people in the community. They still disbelieve, like, that the U.S. bombed the minors. the Minab School, where it's been documented and verified

by so many credible authorities that the U.S. indeed was responsible for the killing of innocent children in that bombing. So it is this very crazy laboratory where we do see the AI accusations just leading these echo chambers into overdrive. I think critical to why we have the information environment,

we have with this war, too, is a, 20 years ago when America went to war in Iraq and Afghanistan, it embedded journalists, you know,

through all the faults with that program, journalists were there documenting everything that was going on. We don't have that now. And, B, Iran has been very good at, we're very effective. at blacking out the internet and cutting off everyone from the wider world, right?

Can you walk us through, like, how the Iranian regime controls the internet, what the blackouts are like. And I'm also, if you can get into it a little bit like the history of it, like how long has this been going on? Was this, when they saw the internet coming, did they kind of set it up this way? Yeah. So, I mean, even before the internet, you have a regime that kind of took power. and really consolidated control over communications because they understood. I mean, the reason why the former regime was able to come into power was because of, you know, the success of like communication and bringing Khomeini's like tapes and smuggle them into the country

and create that social movement. So they really learned the lessons from the fall of the previous regime and they very quickly centralized so many different kind of from the media to like propaganda institutions. So everything was kind of set in place even before the internet to centralize communications. And of course, when the internet came, it was really a threat to the monopoly that they had. And of course, you know, Iranian has a population of over 90 million people. And it's very hard to say what, like, what Iranians think across this massive population. And of course, we can't say because there's no free polling, there's no free elections, there isn't free media, so we can't actually capture the

ratios of what people are saying. And the regime has always tried to present, you know, the one true voice that they hold the legitimacy and desires and they're the representatives of the Iranian people. And of course, there's different indications throughout the years. that this is not true, right? Like, since 2017, on average every two years, there has been a national uprising, and each time those uprising have gotten bigger. So this is one indication that maybe the majority of Iranians are not on board with this regime, right?

And of course, what we see in how they structure the communication is this fear in letting the real authentic voices of Iranians easily come to the fore. and we really saw the power of the internet of social media back in 2009. So 2009, the internet was around. It was the very beginnings of it. It was introduced in the kind of early 2000s in Iran. It hadn't quite become centralized and controlled. And so you had this very symbolic moment where you had this national uprising against,

it was called the Green Movement back in 2009. And you had this viral video of, Neda Arasoltan, a well-known protester, being shot and killed by security forces. So this was one of the first instances that we see in real time the violence of the state, and it was quickly shared on social media, very powerful image. And I think the regime quickly understood the power that the internet brought to allowing people to witness what they do on the ground. Following 2009, obviously there was a violent repression of that movement. and they quickly move to centralize the institutions and the governance and the infrastructure of the internet.

And basically we've seen, and in 2019 we saw the kind of first nationwide shutdown, which was followed by widespread massacres, again, of protesters. And we saw this during the war, the June 2025 war, they also shut down the internet back then. what was interesting was during the June 2025 war, they made the excuse that they were shutting down the internet for security reasons, that there were cyber attacks. I mean, if you looked at the infrastructure, you could see that those cyber attacks were not really related to the civilian connections

that were available. Everyone knew they were afraid of an uprising. They were afraid of losing control. That's why they shut down the internet to stop mobilization. Now, with this current shutdown that is still in its 68th day, following the beginning of the war on February 28th, the kind of veil drop, they weren't making that excuse anymore that this was for national security. They were outright saying that they were blocking the internet and allowing privileged access for those who would carry the message of the state. which was this interesting moment of saying,

this isn't really about access. This isn't about cybersecurity. This is actually about our war propaganda. We want to ensure that there is one cohesive message coming out of Iran. And if you know Iranians, and it was interesting on the first day of the war, on February 28th, when news came out that Khomeini, the Supreme Leader, was assassinated.

There was one state message about that assassination. like a grievous tragedy, unjust strategy. That's the state message. But like, I remember the internet wasn't, didn't shut down until very late on February 28th. And I was in touch with people that I know, I won't say exactly who for security reasons, but they were showing me videos outside of their window of people celebrating. Now, I'm not saying every single person in Iran was celebrating.

Again, 90 million people can't be representative. But a lot of people had that reaction. And there are a lot of diverse perspectives on where people stood on the war. For good or for bad, however you want to judge them. But of course, the regime wanted the one coherent message that this is an unjust war, that the U.S. and Israel are oppressors, and they are standing up against them. Again, this is the narrative that has a lot of buy-in, right? with leftists and the West, with Global South Solidarity,

and I think they really have been effective. And they've been effective in creating that messaging, creating the AI Lego propaganda videos, and they've been effective in blocking out any kind of opposition to their narrative because you can't really attain the real voices of Iranians who would show a very complex picture of what people think. about what's going on about the war. Put a pin in Legos, because I do want to get to that here in just a minute.

There is also on top of the intranet or next to it. There's also a state run intranet too, right? What is kind of, what is the use of that? Does anyone use it? Is it effective? Like, what is that even for? Yeah. Actually, more and more people that I know are using these national messengers on the Iranian internet and people who have been, who have sworn never to buy into these apps or technologies just because there's very little option, right?

You've got like a two-tier system, basically. There is like the open internet where, as you said, if you're if you're like super bonded to the message and they can trust you to carry that message outwards, like maybe you'll get access. And then there's a like an internal network for everybody else, right? Yeah. So it's like the websites, the applications that are built on internal servers that they can control. I mean, it's a very kind of complex to just say it's like two different networks. It's kind of compartmentalize the things that have been, you know, that are domestically located, basically, including things like there's this popular messenger called the Bala app. It's kind of a, like a WhatsApp that they've created. There's obviously

banking infrastructure that's internal. They have, you know, e-medical services that are internal. Obviously, like, these are essential services that's even under whatever circumstances you want to be accessible. And of course, then there's like the messaging, the messaging applications or the search engines and the social media. And like Iranians obviously have to use the banking software and the e-medical services and other social service technologies that are located on the intranet. But there has always been this kind of wariness towards the messaging,

the social media, any other of these kind of like social communications tools that they've had. And they've never successfully been able to get buy-in into these spaces. and, you know, people have spent a lot of money to get VPNs that work. They spend money on expensive international data packages, etc. But with the shutdowns, there has been very little choice. And so people who have wanted to be in touch with their families, initially they were using, you know, phone lines. But even the phone lines are super expensive.

And it's only outgoing calls from Iran to get very expensive international calling cards to make those cards. So I've been seeing more and more people download the Bella app. And of course, this is monitored, surveilled, and like, you can't do anything politically sensitive over this app, but it's cheaper and a way for you to actually check in with your family. So this, like, this shutdown has been effective in actually getting people to buy into the intranet in ways that they were never really able to get people to buy into. When your country is being bombed and it's kind of like siege mode, you're desperate to communicate with everybody no matter what the costs may be, right? Here I'm not talking about monetary costs, obviously. But like, even if you're being surveilled, at least you're able to connect. And that's better than not being able to connect at all.

Exactly. Yeah. I mean, I've had conversations with people. I mean, I was talking to someone, and they're like, well, my brother is in the last stages of cancer. I'm in the middle of the war, and I got like a very questionable VPN that probably is malware, but I don't care because I just want to, you know, be in touch with my brother, who's potentially at the end of his life. So, yeah, people become desperate in these other layers of security and privacy and my rights.

It's like, we don't care about that. It's just these kind of assessment. things, which is how you allow for rights to be eroded and for the state to, I guess, encroach into your private space very effectively. Is that, I assume that that's all within Iran. Like, no, somebody outside, like somebody in the diaspora can't get on that app and then communicate with somebody inside? They can.

Yeah, you can. You can register. I mean, I know people who have been registering onto the ballot app on the desktop. You, yeah, it's very, you can't access every component of ballet. Like if you're not registering, like your national ID or your phone number from a local number, you don't get all aspects of it, but you can still access the, like, messaging. Okay. That's even more interesting because then that's a little piece of the, that kind of state-controlled apparatus

spreading out into the rest of the world. Yep, exactly. Speaking of state-controlled apparatus spreading out into the rest of the world, let's talk about Legos. I think been kind of a Western, interestingly a Western obsession are these AI-generated Lego videos set to rap songs. There's been a ton of them. You know, there's one group called Explosive Media that is kind of responsible for the best ones. There's a bunch of knockoffs. I see these and I think two things.

One is while Iran really understands America and two, as you said earlier, as you noted earlier, like to be able to generate that kind of video and get it online as fast as they are, probably not completely disconnected from the state. And I believe the gentleman behind it has said that he's working on a government contract, right? Yeah, it's really interesting. I remember when I first heard about this.

I was like, they're definitely getting tiered access to the internet in order to be generating content and posting content. I think I gave an interview somewhere and people started attacking me for making that assumption. And then there was an interview they gave. to BBC where they were like, yes, we have tiered access to the internet. The government has given us this privilege. But I think what's really interesting about explosive media, and I think this is something maybe little done, but there's this academic researcher named Narigispojovli. She basically has a book.

It's called Iran Reframed. And the book is based off of... anthropological research she did embedding with besiege media creators in Iran for the course of I think eight years um lots of questions I mean lots of methodological questions of how as an anthropologists you can embed with security forces like the besiege so I don't know if you or your audience knows the besiege are one of the paramilitary units of the Islamic uh republic revolutionary guards, the IRGC. And, yeah, so the besieger, like the most ideological, they're the unit that goes out and represses,

kills protesters. So she had embedded with their media creators to understand how the new generation were going to do propaganda, we're going to kind of represent the ideology of the state. and she actually recently wrote a piece for New York magazine saying that that new generation of besiege media creators that she wrote her book on and did the research on actually explosive media is one of those bassege media creators she studied so they are actually like very much part of the apparatus of the state They are structured in a way where they're kind of like freelancers who have their own kind of consultancies and they sell it like this. But I really don't see the differentiation between explosive media and like officially being part of the apparatus of the state.

And if you have studied propaganda and Iran, if you have studied like influence campaigns, you know it's very much decentralized in these ways where they outsource and sell them. but like the people that they are hiring on these freelance basis, they are essentially part of the state apparatus. Yeah, you make sure that somebody's ideologically sound and then kind of cut them loose to do the work, right? Without oversight, it's, you know, a classic management technique. Yeah. And I mean, yeah, you could read that New York article.

I don't think a lot of people have made the connection that, yes, this researcher is connected. I mean, she doesn't name explosive media explicitly, but she says one of the besiege media creators, she's like checked in on again, is behind popular AI Lego videos. Yeah, is this in a room with social media, in a room with Iran's social media savants? Yes, I believe that's the one in New York Magazine, right? Yeah, I've not read that. I will, and I will link to it in the show notes as well. Um, also anxieties of power, uh, where's the anxieties of power in the Islamic Republic is a great subtitle for, uh, a book. So I will link to that as well in the show notes, um, for people that kind of want to learn more.

So can you tell me what, what is it if you're a young person in Iran, What is the appeal of the regime? Yeah, that's very hard for me to answer. I really can't connect with that, to be honest. I mean, I've been watching the interviews of the guy behind explosive media. Always in a darkened room with the sunglasses on. Yeah, and I mean, I generally, like, whenever I do kind of try to understand the perspectives of the people who like the regime. it's hard for me.

I mean, like, part of the PhD research that I'm doing is trying to understand how technologists have stayed to kind of help build the kind of national internet and the national infrastructure that enables controls in Iran and trying to understand what it takes for people to like stay and to work with the state. and because there's been so much brain drain. Like most of the people, I mean, even in London, if you talk to a lot of recent kind of startup founders from Iran who have basically left because they just were, had come up to a wall in terms of their inability to work or cooperate with the state. Because eventually if your company gets big enough, you have to have some sort of cooperation with the state. And of course you get into like corruption and not only like the rights violations, but

those other elements that you see. So it's very hard for me to naturally want to, to be able to understand or rationalize. It's some combination I think of profit and self-interest and maybe ideological alignment or ideological brainwashing. It's often very hard for me to understand that perspective. but I can't really objectively speak on it because I am so very biased against what the regime stands for. I mean, I understand how they appeal, right? I grew up in the Western world.

Like, I did my formative education in the Western world in the time of the Iraq War. I have, I work in human rights. I see how, like, our entire international legal. frameworks have failed in situations like Gaza and holding entities like Israel and the U.S. accountable for the crimes they've created. So I very much understand the appeal that the kind of external messaging that the Islamic Republic has. But at the same time, I can also understand and hold the other truth, which is the Islamic

Republic is itself an imperialist power that has been. wreaking havoc in the region and on its most importantly, on its own people for decades now. And so it's very hard for me to reconcile or want to be a fan of any of these kind of actors or conflict actors. But like when you do see how people are trying to align in the fact that some of these officials from the Islamic Republic are being thirsted after. Like American women, it's just beyond, it's like, it's really beyond my ability to understand. And I only hope that they're just ignorant of the terrible atrocities that those officials stand for and have helped enable. You want to know my pet theory? This gets real weird about why that happens.

is that in the West specifically, I think people will worship something if they're not careful. And they create fandoms that get obsessive about, you know, you see this video games with pop culture characters, with movies,

with politicians. And I think it's that same kind of energy. It's the same, Like, it's people fall in love with this social media image of a thing online. And it's kind of a fandom that spirals out of control. And I know that that's kind of, it's kind of, it's extremely gross and bizarre when it is regime figures, whether they're American regime figures or Iranian regime figures, people that are responsible for death, horrifying death. But I think that's like what it is. it's it's fandom essentially.

And they're not thinking too deeply, I don't think. I really hope they're not thinking too deeply, because that would be horrific. Yeah. Well, and I'm going to give people perhaps the benefit of the doubt they don't deserve, that they actually don't know. I don't think a lot of people understand, like, this kind of cycle of uprising and repression that's been going on since the Green Movement in 2009, right? that every couple years that like there's this big this big groundswell of people that take to the streets and then the regime comes in and cracks down and there is it's brutal and horrifying and we America is so American are very concerned about American domestic politics and they filter everything that happens outside of the country through those domestic politics and so if if Trump says a thing is bad it's hard for us to process

that you know what, maybe he's right about this one. Even if he's wrong about the way that the war is being conducted and is being conducted poorly and he's posting all of these these slop pictures that doesn't mean that those eight women aren't in very real danger, right? Those are real people. And you've said several times in this conversation, and I really take this to heart.

And it's like part of how I've tried to live my life is that you have to hold more than one thing in your head at the same time. You know, it's kind of like you have like, a lot of this stuff isn't binary. One isn't good and one isn't bad. Sometimes both things are bad. You know, it is, everything's more nuanced. That's my, my rant about why women are thirsting over generals. Yeah.

But I do think we're living in like a social media, AI. everything kind of environment where everything is becoming flattened. Yes. So let's, it kind of at the end of our time here, we talked about the credentials. I know that early on in the war, X said it would be taking, basically demonetizing people that were sharing like video game footage and claiming it was new war, new war footage, these kinds of things. what else can A, platforms do as we churn through this horrible AI generated age that we're in? And B, what can a person that's scrolling do?

Yeah, I mean, I honestly, I think that the onus on the person scrolling is a bit unfair. I do really think the onus should be on the people with the power and the resources. So it really should be on the platforms, on the model creators, right? And it needs to start very early on in the kind of information pipeline at the layer of the models and how they are putting in place these kind of guardrails, putting in place content credentials at the creation stage. And then how this then gets filtered out on the platform level. And honestly, we've been part of C2PA, the Coalition for Content Provenants and Authenticity that's developing these content credentials.

And their companies are there, but they aren't necessarily investing or putting the resources into it in the way that we want them to. There's been really great research by Indicator, even Washington Post, that shows that the companies that have pledged to enact content credentials are just failing. The labels aren't coming. I mean, I have posted so many pieces of AI content on Instagram that never have any label or any sort of transparency. And so it's clear. And obviously it follows the trend. We've seen the trust and safety teams be dismantled at the very same time that they're needed the most because AI is going into overdrive with sophistication and scale. And so it's really not keeping up, and the technology is advancing far faster than any of the guardrails.

So there really does need to be more investment by the companies. I mean, actually, I think tomorrow or the day after tomorrow, meta is supposed to react to the oversight board's recommendations on AI and conflict content. It was something that myself and my team at witness were very involved in, especially in pushing for, the content credentials, which did end up in the recommendations that the board is giving to meta. So we're waiting to see, is meta actually going to step up in the next two days in their reaction and actually want to provide any sort of transparency or content moderation that makes sense for the age of AI? So it's really there. I mean, in the absence of them actually taking action, there are some regulations that are emerging that we are watching very closely.

The California AI transparency bills are going to be enforced in August. And so they are actually going to be requiring companies in California to enact content credentials or some sort of provenance standard. And obviously the content credentials are the most widely accepted form. So watching how that will get enforced, how the companies are going to be complying with actual regulation, I think we'll be giving us an indication of how this might be playing out for the rest of the world, because more and more states are coming out with these bills. You have the EU AI Act's code of practice that will be enforced as well over the summer.

So a lot to watch to see how this works in practice. If people want to find out what's going on with this war, if they want to keep track of it, they're just a normal person. What is your suggestion to them? Like who, who, who, who, who, who, who, who, who, who, who, who, who, who's, who's, who's, who's, who's, in a decent way. Like, if you want up to date information about what's actually happening, as best as we can figure out now, given everything we've talked about today, uh, with, you know, kind of information blackouts on all sides. Yeah. I have some favorites.

Um, the BBC verified. team, I think, has been doing some really excellent job, has been doing a really excellent job of following things, and there's been particular journalists there. I recommend following Shoyan Sardarizada, because I don't think he sleeps, and he's just, like, debunking content, constantly him. And there's another kind of O-Synth researcher. He's actually a Israeli-O-Synth researcher name Tal Hagen and he's done some really good debunking of yeah fake content things that have been going viral like we often get a lot of escalations to go to our AI forensics specialists in the deep fake rapid response force and before we actually turn to um using like

valuable forensics time we we try to look at what Cheyenne and Tal are doing because they do very good objective analysis and debunking of a lot of, or verification of real content, which I would definitely recommend. So there are some, I think, some trusted people doing good OSINTH verification, fact-checking work that should be turned to. And, yeah, more resources and funding for fact-checkers and getting them back onto the platforms, I think is one of the first steps that we need. where can people follow your work?

I am on LinkedIn and Blue Sky. I think most of my updates these days come on LinkedIn for good or for bad. Well, we will, do you want me to post a link to your LinkedIn? Is that? That seems so lame, but sure. I feel so lame to say that I'm on LinkedIn, but such is, the way of the world after a Twitter tanked.

It's just it's bizarre. The vibes are rancid over there and it's so much of it is just the AI generated Lego videos and God knows what else. It's really hard to parse. And the algorithmic

recommendation engine is very strong and aggressive and unpleasant. Yeah. I'm trying to spend more time on blue sky. I'll make more of an effort for blue sky. It's a good community there. most of the time. There's some scolding going on, but it's, you know, it's better than Twitter, I think. Yes, definitely, definitely. Thank you so much for coming on to Angry Planet and walking us through this.

Yes, thank you. I'm going to reach out to some other people you mentioned in this, maybe get some other episodes going. Thank you so much. Amazing. Yeah. Thanks so much for having me. Looking forward to hearing this. That's all for this episode of Angry Planet. As always, Angry Planet, is myself in Kevin Nodell. It was created about myself and Jason Fields. If you like the show, angry planetpod.com,

you get life during Trump time. You get bonus episodes of the show. You get early access to all the mainline episodes. I know we took a little bit of a pause there. I had pneumonia. I'm sorry. It's a bit of a one-and-a-half-man operation here. And things are getting a little hairy.

Thank you for sticking with me. And we will be back again soon with another conversation about conflict on an angry planet. Stay safe until then.