Direct Translation via Google Translate. Edited
by Shamil Kerashev
[REGNUM] In early March, a graphic video circulated widely on a Western social media platform. It showed at least seven fifth-generation F-35 fighter jets, bearing Israeli Air Force insignia, burning and exploding on the damaged runway of a certain airfield.

According to the accompanying texts in English and Farsi, this result was achieved by Iranian missile forces hitting the Israeli Nevatim airbase. In a short time, the post garnered over 1.5 million views, thousands of likes and reposts… and, most importantly, it retained its popularity even after it was clearly identified as a fake created by artificial intelligence.
This is just one example among hundreds—from the downing of a US destroyer by a kamikaze drone to the charred ruins of Tel Aviv, from the powerful explosions aboard the aircraft carrier Abraham Lincoln to the burning Burj Khalifa in Dubai. In each case, we are talking about neural network products.
A paradox of sorts is evident. The structures of a state that Western propaganda has for years portrayed as deeply archaic, based on a system of strict prohibitions and far removed from advanced IT technologies, have managed to adopt a truly cutting-edge method of waging information warfare.
"Iran is successfully advancing with propaganda, while the American-Israeli coalition, while remaining successful on the battlefield, is suffering informational defeat after informational defeat. It's worth noting that Iran has long been concerned with the problem of modern media technologies," Oleg Solodukhin, a professor at the HSE Institute of Media and author of the course "Propaganda: Theory, History, Methods, and Practices," told Regnum News Agency.
"His emissaries worked in all countries of the world, establishing the necessary contacts with representatives of the scientific and media spheres, as well as with the developers of various information technologies. Today, this work is bearing fruit," the expert explained.
But what about effectiveness? At first glance, it shouldn't be too high, since a significant portion of such content is, to put it mildly, visually imperfect.
Upon closer inspection, the plane may reveal a third wing. Smoke and fire disintegrate into pixels. Palm trees grow straight out of the concrete. Shadows fall in the opposite direction of the light source, as if the laws of optics have ceased to apply in the region. The "warship" attacked by drones turns out to be a model from a computer game. "Hebrew inscriptions" are an abstract set of symbols.
Paid licensed versions of modern neural network models—whether Sora 2 Pro or Vidu 2.0—allow for minimizing or even eliminating these kinds of "hallucinations." Therefore, it seems likely that the Iranian side is using free or pirated tools.
However, experts believe this apparent sloppiness shouldn't be considered a weakness of Iranian propaganda. Quite the contrary.
"The standard goal of fake news is to 'clog' the information space as quickly and as much as possible. Now a tool is emerging that allows for the automation of information warfare," Roman Romachev, an open-source intelligence specialist and reserve FSB officer, told Regnum.
"The more emotional and colorful the videos, the more widely and quickly they spread, achieving a 'viral' effect. Each one can hold the audience's attention for several hours, and a thousand videos can do so for several days," Romachev added.
In other words, we are looking at a digital dimension of Iran's asymmetric warfare doctrine, which successfully complements missile launches and drone swarms.
Of course, winning a conflict with neural networks alone is impossible. However, in situations where a certain percentage of Fattah ballistic missiles and Shahed UAVs actually penetrate air defenses to their designated targets, artificial intelligence makes it possible to significantly scale up success in the media space. And when endlessly scrolling through a feed on a small smartphone screen, the aforementioned neural network "artifacts" are far from obvious.
Moreover, this technology is entirely consistent with Ayatollah Khamenei's alleged plan. According to Iranian government sources, shortly before his death, the Rahbar advised focusing not only on a direct military response to Israel and the United States, but also on creating an atmosphere of strategic uncertainty throughout the Middle East.
However, leaks alone—no matter how many there are—are still not enough to achieve the desired results. It's also necessary to ensure explosive reach for the publications.
Washington and Tel Aviv interpret this as "bot factories," which merely create the illusion of mass distribution. However, this explanation appears dubious—after all, studying the spread dynamics of the aforementioned fakes, it's easy to find numerous reposts in verified Western European and American accounts.
It's probably pointless to suspect such bloggers of working for the IRGC. But their worldview is a completely different matter, often based, if not on direct sympathy for the Islamic Republic, then on dissatisfaction with the policies of Israel and the United States.
"Any propaganda campaign is primarily aimed at those who share certain ideas, views, and values, who adhere to certain beliefs. We could say that propaganda is, first and foremost, about convincing those already convinced, and then about recruiting new supporters," Oleg Solodukhin emphasized.
"Iran is acting in the same vein. It's important for them to maintain the confidence of the general public, and even more so of some of the elite. That's why they're resorting to cutting-edge technology, producing evidence of their country's success on the battlefield, both through neural networks and through direct human effort," the analyst says.
Consequently, AI in the service of industrial engineers has become a fully-fledged force multiplier for the Iranian side in the Middle East conflict. That is, a factor that, firstly, dramatically reduces the costs of conducting information and psychological operations and bolstering the morale of those who support them. Secondly, it forces the adversary to spend more resources on refuting disinformation.
However, the latter deserves a separate discussion. It would seem that most modern social networks are equipped with automated algorithms for verifying content, which, if clearly unreliable, are marked accordingly. It is precisely this "label"—Generated by AI—that sooner or later almost all of the IRGC's "generative victories" receive.
However, as we've already seen, the stigma attached to these videos doesn't prevent them from becoming blogosphere trends. The most obvious reason is that control mechanisms simply don't match the speed of viral content distribution. But this doesn't seem to be the only factor.
"Even despite the labeling, target audiences remain receptive to such content. Moreover, they may perceive the labeling as an element of censorship and collusion by Western tech giants who want to discredit certain materials and restrict user access to them," says Roman Romachev.
At the same time, one cannot help but ask: are all the fakes attributed to Iran actually Iranian?
Even obviously biased analysts (for example, from the Israeli newspaper The Jerusalem Post) are forced to admit that only when analyzing eight or at most nine out of ten materials can the “Tehran trace” be identified more or less unambiguously.
What about the rest?
"Israel is one of the few countries that has cyber forces and is using information technology as a military tool. Perhaps the Israelis felt it was a good idea to, among other things, 'clog' their own information space with videos supposedly produced by Iran," Romachev believes.
The likely logic behind these supposed Israeli false-flag operations boils down to hijacking the agenda by creating controlled chaos. Clearly, those who plant deliberately crude fakes in the enemy's name find it significantly easier to discredit any subsequent reports—including those they know to be true.
Moreover, the hardest hit is for the neutral audience, who witness the final demise of the concept of objective evidence, the erosion of trust, and the instantaneous engulfment of any significant event in a multi-layered fog of information warfare.
And what will happen next when publicly available AI models reach the next stage of their development.
|