Identifying, Tracking and Locating Shadow Vessels in the Age of Artificial Intelligence

The growth of maritime trade, international sanctions regimes and clandestine logistics has given rise to a new category of elusive ships often called “shadow vessels”. These are tankers, bulk carriers or smaller craft that operate covertly to evade detection, frequently involved in smuggling, sanctions-busting, or illicit military resupply. Such vessels adopt deceptive practices: turning off their Automatic Identification System (AIS) transponders, “spoofing” location signals, assuming false identities, changing flags of convenience, or conducting ship-to-ship transfers at sea. The challenge of identifying, tracking and locating them has become a central concern for naval authorities, coast guards, insurance firms and intelligence services. Modern artificial intelligence (AI), combined with satellite surveillance, signals intelligence, and predictive analytics, has transformed this field.

Exploiting AIS Anomalies with AI

Under the International Maritime Organization’s rules, most vessels above a certain size must broadcast their position, speed, heading, and identity via the AIS. Shadow vessels routinely manipulate or deactivate their AIS signals to disappear from public tracking platforms. Modern AI systems can detect anomalies in AIS data even before a vessel disappears entirely. Algorithms compare transmitted positions with historical routes, fuel capacity, declared destinations and maritime weather patterns. Machine learning models trained on millions of normal voyages can highlight suspicious behaviours such as abrupt course changes, improbable speeds, or signals appearing simultaneously from geographically separate locations (“ghosting”). Even when AIS is turned off, the time and location of last transmission can help predict where the vessel might next emerge.

Synthetic Aperture Radar and Multispectral Satellites

Where ships attempt to hide from optical surveillance under cloud cover or in darkness, synthetic aperture radar (SAR) satellites offer persistent, all-weather imaging. SAR can detect the metal hulls of ships regardless of lighting, making it ideal for monitoring oceans where AIS coverage is weak. Multispectral and hyperspectral satellites add another layer, allowing analysts to pick out oil slicks from covert ship-to-ship transfers, or unusual heat signatures from engines running at night. AI now processes this deluge of satellite imagery at scale, automatically classifying ship shapes, measuring wakes to estimate speed, and cross-referencing them with known vessel registries. This reduces what was once a painstaking manual process to near-real-time detection.

Signals Intelligence and Passive Emitter Tracking

Even stealthy vessels rarely operate in complete radio silence. Communications with shore bases, encrypted satellite phones, or the use of radar and navigation equipment create a spectrum “fingerprint”. AI-assisted signals intelligence (SIGINT) platforms can intercept and classify these emissions. By triangulating weak signals from multiple receivers—whether drones, aircraft, or shore stations—authorities can approximate a vessel’s location without it ever broadcasting AIS. With machine learning, these spectral fingerprints can be matched to known equipment fits or shipping lines, narrowing down which specific ship is being tracked.

Drone and Aircraft Patrols with Computer Vision

Maritime patrol aircraft and unmanned aerial vehicles (UAVs) can sweep large ocean areas. Equipped with electro-optical cameras, infrared sensors and even miniature SAR pods, they provide a flexible and mobile layer of surveillance. Modern computer vision algorithms process live video feeds, recognising hull shapes, deck equipment or paint schemes associated with sanctioned fleets. They can even match tiny features like lifeboat placement or antenna masts against a database of shipyard blueprints, enabling a positive identification even after a vessel has been repainted or renamed.

Predictive Analytics and Behavioural Modelling

Once sufficient data on a vessel’s movements exist, AI can be used not merely to follow her but to anticipate her behaviour. Predictive models incorporate historical routes, cargo types, seasonal demand, port access records, and even weather forecasts to estimate the most likely rendezvous points for ship-to-ship transfers or illicit port calls. This permits pre-positioning of assets for interception or imagery collection, increasing the probability of catching the vessel in the act. Insurance companies and banks also use such models to identify high-risk vessels before issuing cover or financing.

Commercial and Open-Source Collaboration

The proliferation of commercial satellite constellations, open maritime databases and crowd-sourced intelligence has democratised ship tracking. Networks of amateur observers, maritime photographers and AIS hobbyists upload data that can be fused with government intelligence. AI platforms ingest this heterogeneous information, correlate it with restricted databases and produce risk scores for each vessel. This blending of open-source intelligence (OSINT) and classified feeds is a hallmark of modern shadow vessel detection.

Emerging Technologies: Quantum Sensors and Undersea Networks

The future may bring even more intrusive methods. Quantum navigation sensors, able to detect subtle anomalies in Earth’s magnetic field, could reveal submerged or uncooperative vessels without reliance on satellite signals. Distributed networks of autonomous underwater vehicles (AUVs) may patrol choke points, using acoustic signature recognition to track hulls passing overhead. These technologies are in development but already demonstrate the trajectory: towards persistent, multi-domain maritime awareness where few ships can hide for long.

Legal and Diplomatic Constraints

While the technology for identifying and tracking shadow vessels is advancing rapidly, the law of the sea imposes strict limits upon what states may do with this information. The United Nations Convention on the Law of the Sea (UNCLOS) divides the world’s waters into territorial seas, exclusive economic zones (EEZs), and the high seas, each with differing degrees of jurisdiction. Within territorial seas, coastal states enjoy broad enforcement rights, including inspection, seizure and arrest. Within EEZs, rights are limited to resource management and environmental protection, not general policing. On the high seas, vessels are subject solely to the jurisdiction of their flag state, save for a few exceptions such as piracy, slavery, and unauthorised broadcasting.

This means that even when a shadow vessel is identified by AI analysis, a coastal state cannot necessarily act unless it can demonstrate a breach of international law that confers jurisdiction. The doctrine of “exclusive flag state jurisdiction” often shields these ships, particularly those sailing under so-called flags of convenience from permissive registries. Boarding a ship without the flag state’s consent risks a diplomatic incident or even a breach of international law. Consequently the intelligence gained through advanced tracking is frequently used for diplomatic pressure, financial sanctions, or denial of port access rather than direct interception at sea.

Additionally the use of surveillance satellites and remote sensing technologies over the high seas is not explicitly regulated by UNCLOS but may raise questions of sovereignty and privacy. Some nations object to persistent foreign monitoring of their fleets, considering it quasi-military activity. Therefore states employing AI-based detection systems often rely upon multilateral frameworks, such as the International Maritime Organization, the Financial Action Task Force, or United Nations sanctions committees, to legitimise and share their findings. This diffusion of enforcement through cooperative diplomacy rather than unilateral naval power defines the contemporary effort against shadow fleets.

Case Studies: Iran, Russia and North Korea

The Iranian Oil Fleet

Following decades of sanctions on her petroleum exports, Iran has developed a sophisticated network of shadow tankers operating under multiple identities. These vessels routinely disable AIS transponders in the Gulf and conduct ship-to-ship transfers in the East China Sea or off the coast of Oman. Satellite analysis by Western intelligence agencies and private firms such as TankerTrackers and Windward uses SAR imagery to detect the tell-tale patterns of oil transfers—two tankers stationary and beam-to-beam for hours in open waters. AI correlates this with oil production data and port departure times in Kharg Island (an Iranian island in the Persian Gulf), allowing estimation of export volumes even without direct visibility. Iranian vessels often use false flag registrations from countries such as Panama, Liberia or Tanzania, and sometimes repaint names between voyages. By comparing satellite imagery to hull blueprint databases, analysts have unmasked dozens of ships masquerading as others, tightening the noose on Iran’s sanctions evasion network.

Russia’s Sanctions Evasion Fleet

Since 2022, Russia has assembled an ever larger shadow fleet to transport crude oil outside the G7 price cap mechanism. These vessels are typically old, uninsured tankers purchased through opaque intermediaries and flagged in jurisdictions with weak oversight. They frequently conduct mid-ocean transfers near Ceuta (Spanish Morocco), Kalamata (in southern Greece) or the Laconian Gulf (another remote area of southern Greece) - all in the Mediterranean Sea. European and American analysts employ AI-enhanced maritime domain awareness systems to monitor these “dark activities”. By integrating radar satellite data, AIS anomalies and optical imagery, they can infer the flow of Russian oil despite efforts to obscure it. The EU Maritime Analysis and Operations Centre (Narcotics) and NATO Maritime Command have both adapted counter-narcotics detection algorithms to identify likely ship-to-ship transfers associated with Russian exports. The result has been targeted insurance restrictions, port bans, and seizure of tankers in European harbours.

North Korean Sanctions Evasion

North Korea, under comprehensive United Nations sanctions, has long relied upon covert maritime operations to import fuel and export coal. Her vessels often broadcast false AIS identities or use “spoofed” GPS signals showing them anchored in Chinese ports while in fact operating offshore. AI-assisted pattern recognition of radar imagery has enabled analysts to identify North Korean hulls even when their names are changed. Ship-to-ship transfers in the Yellow Sea are now routinely captured by commercial satellites, with data shared through the United Nations Panel of Experts. These images form the evidentiary basis for sanction enforcement and diplomatic démarches. In some instances, signals intelligence has confirmed communications between North Korean captains and Chinese brokers, illustrating how digital intercepts can complement visual surveillance.

Conclusion

In the past a ship could vanish beyond the horizon, protected by the opacity of the sea. Today’s maritime domain awareness, powered by AI, satellite constellations and sensor fusion has eroded that invisibility. No single method guarantees success—shadow vessels adapt constantly, refining their disguises and routes—but the combination of AIS anomaly detection, SAR imagery, SIGINT, aerial surveillance and predictive modelling creates a multi-layered net. Yet technological mastery must coexist with legal restraint: the seas remain a global commons, and enforcement depends as much upon diplomacy as upon data. The case studies of Iran, Russia and North Korea reveal the evolving cat-and-mouse game between those who conceal maritime trade and those who illuminate it. The oceans are becoming transparent, but sovereignty still casts its long shadow.