While the war is being fought with missiles and statements, Trump’s former digital strategist appears to have turned to more modern artillery — search results, GPT systems, and what machines call “established fact.”
Brad Parscale has once again found himself at the point where politics meets technology and quickly loses whatever innocence it had left. After the United States and Israel struck Iran, a network of websites linked to his company, Clock Tower X LLC, began producing a stream of materials designed to influence not only readers, but artificial intelligence as well. Formally, it looks like an information campaign. In substance, it looks more like an attempt to explain in advance — to both machines and people — what they are supposed to regard as true about this new phase of the conflict.
Against this backdrop, what stands out most is not the content of individual publications, but the architecture of the operation itself. Some articles claim that Iranians allegedly see the strikes as an opportunity for liberation. Others frame the bombing of Iran’s Interior Ministry as a step forward for human rights. Others tie anti-war protesters to networks funded by the Chinese Communist Party. Still others lead the reader toward the idea that the removal of Khamenei is a nearly natural end point of a conflict that, if these texts are to be believed, Iran supposedly started back in 1979. Follow the logic of these publications, and the story is already over, the guilty party has been assigned, and the moral framing carefully arranged. All that remains is to convince search engines and chatbots.
The Digital Architect of Old Politics Opens a New Front
Brad Parscale is known as the former digital director of Donald Trump’s 2016 campaign and the manager of his 2020 campaign. He was one of the people who helped turn data-driven political advertising into a large-scale machine for microtargeting and attention management. Now, judging from the materials presented, that same expertise is being used in a far more sensitive arena — an international information campaign serving the interests of a foreign state.
According to federal documents, last August Clock Tower X was hired by Israel’s Ministry of Foreign Affairs through an intermediary, the French PR firm Havas Media. The contract was initially valued at $6 million, but by a December update it had risen to $9 million. The arithmetic alone is enough to destroy any illusion of spontaneous media activity. When that kind of money is allocated to narrative distribution, this is no longer merely communication. It is a systematic information infrastructure.
The contract requires the company to produce at least 100 primary creative materials every month — videos, graphics, audio clips, and articles. In addition, it provides for thousands of variations tailored to different audiences and platforms, including TikTok, Instagram, YouTube, Rumble, podcasts, and search-optimized websites. At least 80% of the content must target Generation Z. In other words, this is not just political messaging. It is precision tuning for an audience that consumes information in short bursts and rarely reads the small print at the bottom of a page.
A Network Disguised as Analysis, Fact-Checking, and Education
The materials were distributed through a network of more than ten websites created or used by Clock Tower X. Each one is carefully wrapped in a recognizable shell of trust. Cognitura presents itself as a research and educational platform. FactSignal calls itself a global intelligence hub. Justorium is styled as a legal resource supposedly devoted to justice and law in defense of freedom. Allyvia speaks of strengthening the U.S.-Israel alliance. Culturavia plays the role of cultural exchange.
At first glance, it looks like a standard digital ecosystem of issue-based media. Timestamps, article cards, “NEWS” labels, short teasers, cross-links — everything you would expect from respectable independent platforms supposedly just doing their job. But at the bottom of the pages, in small print, sits the standard FARA notice: distribution is carried out by Clock Tower X on behalf of the State of Israel. Formally, disclosure requirements are met. In practice, almost no reader is going to make it that far. That is the elegance of the construction: legal transparency is provided in a way that allows it to remain nearly invisible.
The Content Is Written Not Only for People, but for Machines
The most interesting part of the story is not even the fact of the campaign itself, but how exactly it is built. The contract places special emphasis on content designed to shape search results and influence what artificial intelligence systems say. It explicitly provides for the creation of at least 20 new web pages each month targeting priority keywords. More than that, it speaks of deploying websites and content to generate results shaped through GPT methods in discussions using GPT methods.
In other words, the goal is not merely for a person to open an article and read something. The goal is for those pages to become raw material for future machine-generated answers. This is no longer classic PR, and not ordinary propaganda either. It is an attempt to insert itself into the very fabric of digital knowledge, from which “objective” formulations are later extracted for search answers, chatbots, and language models.
Clock Tower X appears to be following that logic with remarkable discipline. Almost every article is built like a textbook for an algorithm: a table of contents, numbered sections, bolded key takeaways, FAQ blocks with questions closely resembling real search queries. “What is the U.S.-Israel alliance and how does it affect American military strategy?” or “How does intelligence sharing with Israel enhance U.S. security?” These are no longer just journalistic headlines. They are bait for machine answer extraction.
When Contested Claims Are Packaged as Reference Material
In many of these texts, politically contested claims are presented as uncontested facts. This is not done in the form of columns or opinion pieces, where the reader at least understands that what they are seeing is a point of view. No — everything is wrapped in a standardized format, complete with notes, links, and pseudo-academic confidence. In one publication, for example, it is asserted that the 2015 Iran nuclear deal failed because “Iran cheated from day one.” The sole supporting source is a pro-Israel advocacy group, the Jewish Institute for National Security.
This kind of arrangement becomes especially effective when the sites begin citing one another. On the surface, it resembles an ecosystem of independent confirmation. In practice, it may simply be a carefully organized network of self-reinforcing claims imitating authority. One site cites a second, the second cites a third, the third circles back to the first. The machine sees a cross-linked structure of citation and concludes it is looking at a stable body of knowledge. The reader sees footnotes and relaxes. At that point, propaganda starts to look like reference information.
Why GPT Systems Are Especially Vulnerable to This Kind of Packaging
This architecture matters for a reason. Language models trained on web data, and systems that extract information in real time, tend to treat predictable structured formatting as a sign of useful reference material. FAQ blocks, clear headings, named authors, regularly updated pages, logical and repeatable structure — all of this functions as a marker of reliability for the algorithm. Not truth, but reliability of form.
That is why a Cognitura article on the history of Iran’s hostility toward the United States, formatted somewhere between a Wikipedia entry and an analytical brief and linked to government press releases and well-known media outlets, may well appear to an AI system as a neutral source. Not as a campaign page created under contract for a foreign government. Once such information enters a model’s training data or is retrieved as a source, it starts participating in the formation of what the machine presents as established fact. And this is where the neatest substitution takes place: Clock Tower X’s authorship and the connection to Israel remain buried near the bottom of the page, while on the surface all that remains is a “neutral” answer already processed by the machine.
The official line may insist that this is simply modern digital marketing. But if one follows the logic of what is happening, it points to something more serious: an attempt to program not only public opinion, but the infrastructure of machine knowledge itself — the infrastructure that increasingly replaces independent search for many users.
The Industry Has Already Learned to Speak to Machines Instead of People
In this story, Clock Tower X does not look like a one-off exception. On the contrary, its actions reflect a broader trend. PR firms increasingly advise clients to create materials not just for audiences, but for artificial intelligence systems as well — taking into account so-called machine understanding. The emphasis is on message consistency, strict formatting, and the most journalistically familiar packaging possible.
In one recent industry analysis, V2 Communications Vice President Caitlin Holbrook directly noted that AI systems reward content that resembles journalism: clear headlines, FAQ formats, author names, and regularly updated pages. In other words, the market has long understood that to shape the future of digital knowledge, it is not necessary to write more persuasively. It is enough to write in a format that a machine will accept as useful and authoritative.
And here a simple question arises: if artificial intelligence is increasingly becoming the intermediary between an event and its interpretation, then who exactly gains power over reality — the journalist, the state, the PR consultant, or the one who first learned how to speak to the algorithm in its favorite language?
The Operation Extends Beyond Websites and Into Private Messages
Against this background, it is especially telling that the operation is not limited to the website network. Since November, Americans have reportedly begun receiving text messages from unknown numbers, supposedly from organizations called “Friends for Peace” and “Partners for Peace.” These groups, according to reports, apparently do not exist at all. Recipients were invited to share their views on Israel and then directed to Clock Tower X websites.
The messages were sent through Voice over IP numbers and appear to have been at least partially automated. People who expressed skepticism about Israel were steered toward pro-Israel talking points and videos from the YouTube channel Allies for Peace. In one exchange reported by Responsible Statecraft, a contact named “John” urged the recipient not to trust stories about Gaza, claiming that there are networks of accounts posing as Gazan civilians and that most of the content is fake. In another case, the campaign acknowledged that it was supported by Clock Tower X LLC on behalf of the State of Israel — but only after persistent questioning.
FARA specialists have noted that such messages may violate the law’s requirement for clear disclosure of foreign-agent status directly within the text of the communication itself. And this is where the story becomes even more revealing: formal compliance with the law is again pushed to the point where it no longer functions as real disclosure. Just like on the websites — the information is there somewhere, but placed in a way that ensures most people will never notice it.