A report in the Financial Times that Russian security services temporarily shut down some of the surveillance systems protecting President Vladimir Putin following the assassination of Iranian Supreme Leader Ayatollah Khamenei highlights how AI-powered intelligence tools are rapidly reshaping global security thinking.
The move follows concerns in Moscow that similar technology used in Iran is said to be helping Israeli intelligence agencies analyze vast amounts of traffic camera footage, potentially exposing vulnerabilities in Russia’s own extensive surveillance infrastructure.
The FT reported that Russian authorities, concerned that a system designed to monitor the country’s population could itself be turned into a tool for foreign intelligence operations, prompted engineers to temporarily disable parts of Putin’s personal surveillance camera network and isolate them from the internet.
This concern is directly related to Iranian operations, with Israeli intelligence reportedly using hacked traffic cameras and AI systems to track movement patterns and determine the timing of high-level meetings in Tehran ahead of the February 28 assassination.
The incident, first reported in early March, marked a turning point in the way intelligence agencies view everyday surveillance systems not just as surveillance tools, but as data-rich networks that can be mined, analyzed and weaponized at scale.
At the center of this change is a new generation of AI companies, including Israeli startups Airis and Conntour, that are building tools designed to make sense of large amounts of fragmented video data.
Airis, which has raised $60 million to date from investors including PSG Equity and TLV Partners, is developing an AI system that integrates input from traffic cameras, drones, body-worn cameras, and other visual sources into a unified intelligence picture. The system is designed to enable real-time querying and analysis of large-scale video data.
The company was founded by veterans of Israel’s intelligence and defense ecosystem and has already been deployed by government agencies and is part of the U.S. Army Technology Program.
Conntour, a Tel Aviv-based startup founded in 2024 and backed by investors including General Catalyst and Y Combinator, takes a complementary approach. Its platform allows users to search video networks using natural language, for example to query for footage of specific objects, individuals, or actions without relying on predefined categories.
Both companies reflect a broader shift in surveillance technology from static surveillance systems to dynamic, searchable intelligence networks.
The Iran incident became a reference point for this new era.
Israeli intelligence reportedly spent years accessing and analyzing traffic camera footage from across Tehran to develop “patterns of life” in the behavior of security personnel and officials. We combined this analysis with other sources of information to map movements and identify operational windows.
This operation demonstrated how large-scale surveillance systems originally designed for public safety can be repurposed into intelligence tools when combined with modern AI systems that can process vast amounts of unstructured video data.
It also caused international concern. Security agencies in several countries are currently reassessing the resilience of their surveillance infrastructure, particularly transportation networks and urban camera networks that were once considered low-risk.
