How AI Is Reshaping the Information War Around the Iran Conflict
This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here.
"Anyone wanna host a get together in SF and pull this up on a 100 inch TV?"
That post on X was referencing a live intelligence dashboard tracking the US-Israel strikes against Iran in real time. Assembled by two analysts from venture capital firm Andreessen Horowitz, the platform stitches together open-source data streams — satellite imagery, maritime tracking feeds, and geospatial signals — alongside a chat interface, curated news aggregation, and direct links to prediction markets where users can wager on outcomes like the identity of Iran's next supreme leader. The recent confirmation of Mojtaba Khamenei left some participants with a notable payout.
Over the past week, I've reviewed more than a dozen platforms operating in this space. A significant number appear to have been rapidly prototyped — "vibe-coded," in the parlance of the moment — over the course of a few days using AI-assisted development tools. One such platform caught the attention of a co-founder of intelligence giant Palantir, the same platform currently serving as the US military's gateway to AI models like Claude during active operations in the conflict. While some of these dashboards predate the Iran campaign, virtually all are being positioned by their creators as an antidote to sluggish, unreliable mainstream media — a direct line to ground truth. "Just learned more in 30 seconds watching this map than reading or watching any major news network," one user wrote on LinkedIn, responding to a real-time visualization of Iranian airspace closures ahead of the strikes.
Much of the discourse around AI and the Iran conflict has, justifiably, centered on the role that models like Claude may be playing in informing US military targeting decisions. But these intelligence dashboards and the broader ecosystem surrounding them illuminate a separate, equally consequential function AI is now performing in wartime: shaping the information environment — and frequently distorting it.
Several converging forces are driving this phenomenon. AI-powered development tools have dramatically lowered the barrier to assembling open-source intelligence pipelines, while large language models offer rapid — if often unreliable — analysis of incoming data. A proliferation of synthetic and manipulated media has left conflict observers hungry for the kind of raw, verified intelligence traditionally accessible only to state-level agencies. Demand is further amplified by real-time prediction markets that create direct financial incentives for staying ahead of the news cycle. And the widely reported deployment of Anthropic's Claude by the US military — despite its classification as a supply chain risk — has functioned as a powerful signal to the broader public that AI is the instrument of choice for serious intelligence work. The cumulative effect is a new category of AI-enabled wartime spectacle, one that can obscure understanding as readily as it advances it.
As a journalist, I see genuine potential in these tools. The ability to surface real-time data on shipping disruptions, infrastructure outages, or airspace closures in a single unified interface is genuinely powerful. But consuming a live war through that lens — snacks in hand, bets placed — transforms a humanitarian catastrophe into a form of interactive entertainment. More critically, there are substantive reasons to question whether raw data aggregation produces the kind of understanding it appears to offer.
Craig Silverman, a digital investigations expert who teaches open-source investigative techniques, has been cataloguing these platforms — his count now stands at 20. "The concern," he explains, "is there's an illusion of being on top of things and being in control, where all you're really doing is just pulling in a ton of signals and not necessarily understanding what you're seeing, or being able to pull out true insights from it."
Signal quality is a foundational problem. Most dashboards feature AI-generated summaries of rapidly evolving, high-complexity news events — a pipeline that reliably introduces errors and omissions. The underlying design philosophy prioritizes volume over curation, presenting strike coordinates alongside obscure cryptocurrency price tickers with no meaningful hierarchy or editorial judgment applied.
Professional intelligence operations work differently. Analysts pair raw data feeds with domain expertise, historical context, and institutional knowledge built over years. They also draw on classified and proprietary sources that simply don't exist in the open-source layer these dashboards are scraping.
The implicit pitch from the builders and distributors of these platforms is one of democratization: that AI can finally break the elite monopoly on high-quality intelligence and deliver it to anyone with a browser tab, whether their goal is to stay informed or to place bets on nuclear strikes. But information volume — something AI is genuinely adept at generating — is not a substitute for accuracy or analytical context. Intelligence agencies solve this problem internally. Rigorous journalism has historically done the same for the public. Neither function is replicated by a dashboard.
The entanglement with prediction markets deserves particular attention. The Andreessen Horowitz dashboard features a live feed of wagers being placed on Kalshi — a platform the firm has invested in. Other dashboards route users to Polymarket, with active markets on questions like whether the US will extend strikes to Iraq or when Iranian internet connectivity will be restored.
AI has also continued to accelerate the production and distribution of synthetic content, and the Iran conflict has become a vivid case study in that failure mode. Last week, the Financial Times documented a wave of AI-generated satellite imagery circulating across social platforms.
"The emergence of manipulated or outright fake satellite imagery is really concerning," Silverman says. Satellite imagery carries an inherent credibility premium with general audiences — it reads as objective, technical, and authoritative. Systematic forgery of that medium threatens to undermine one of the most reliable evidentiary tools available for documenting what is actually happening on the ground.
What we're left with is a sprawling ecosystem of AI-generated content — dashboards, derivative betting markets, imagery both authentic and fabricated — that makes this conflict measurably harder, not easier, to understand.
