# Kremlin disconnects cameras after AI-enabled espionage scare > Source: > Published: 2026-06-15 07:42:54.133095+00:00 # Kremlin disconnects cameras after AI-enabled espionage scare The Financial Times reported that Russia's security services temporarily disconnected parts of a special surveillance system protecting President Vladimir Putin and his close circle after the US-Israeli killing of Iran's supreme leader, Ayatollah Ali Khamenei, according to France 24's coverage of the FT story. The FT wrote that Israeli intelligence harvested hours of Iranian traffic camera footage and used artificial intelligence to process large volumes of video, allowing analysts to pinpoint timing and location. France 24 interviewed surveillance expert Ksenia Ermoshina, who said internet shutdowns and other measures have been used in Russia around political events and that AI video analysis raises new security concerns for the Kremlin. ### What happened The Financial Times reported that Russia's security services temporarily disconnected parts of a special surveillance system that protects President **Vladimir Putin** and his close circle, according to France 24's reporting on the FT story. The FT account attributes the action to concerns that intelligence agencies used hours of traffic camera footage from Iran, processed with **artificial intelligence**, to identify the exact time and location of a meeting involving Iran's supreme leader, **Ayatollah Ali Khamenei**, after his killing by US-Israeli forces, per the FT coverage cited by France 24. The FT described this case as an example of how AI can process millions of hours of video from thousands of cameras to "extract patterns and secrets at an industrial scale," according to the FT excerpt reproduced by France 24. France 24 also published an interview with surveillance expert **Ksenia Ermoshina**, who said, "Depending on political events, big celebrations, events like the St. Petersburg International Economic Forum, or in border regions with military operations, internet shutdowns can be activated without any reason - for one week, two weeks or more." ### Editorial analysis - technical context Industry-pattern observations: modern video-analytics pipelines combine large-scale camera networks, automated object and face recognition, and temporal correlation algorithms to link disparate clips across time and location. The FT reporting highlights that combining vast publicly available camera feeds with AI-powered pattern extraction materially lowers the cost and time needed to reconstruct movements, which amplifies intelligence collection risks in any environment with dense camera coverage. Observed patterns in similar cases show intelligence services can exploit commercially available computer vision tools together with manual analyst review to derive high-confidence location and timing insights from noisy video sources. ### Context and significance For state-level and executive protection contexts, the reported shutdowns indicate operational concern about exfiltration vectors that bypass traditional perimeter defenses, namely the aggregation and AI processing of widely distributed camera footage. For security practitioners and analysts, the incident underscores that the threat surface for high-value targets now includes third-party and foreign camera networks plus open-source imagery that can be stitched and analyzed at scale. For civil-liberties observers, the same technical capabilities that enable surveillance also increase the incentives for tighter internet and camera controls, a dynamic noted in the FT and France 24 pieces. ### What to watch For practitioners: monitor developments in three areas, drawing only on publicly visible signals. First, public reporting and vendor disclosures about increased use of air-gapped or local-only camera architectures in high-security sites. Second, announcements or procurement notices from government protection services referencing camera-network hardening, encryption, or offline storage mandates. Third, advances in adversarial techniques and countermeasures for video-analytics, including policies or toolkits aimed at limiting large-scale footage aggregation. Observers should also watch for additional reporting that cites named intelligence or vendor sources for confirmation of the FT account. ## Scoring Rationale The story is notable for demonstrating a practical risk vector where AI dramatically lowers the cost of analyzing distributed camera footage for intelligence targeting. The Kremlin response - temporarily disconnecting a presidential protection network - shows real operational impact. It is not a model release or regulatory watershed, but matters for security engineering and threat modeling. Practice interview problems based on real data 1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with. [Try 250 free problems](/problems)