Paper ReviewComputer SystemsSystematic Review

Cybersecurity at Sea: AI-Driven Threat Detection for the Maritime Industry

The maritime industry is undergoing rapid digitalizationโ€”autonomous vessels, IoT-connected cargo systems, satellite-dependent navigation. This digital transformation has exposed critical cybersecurity vulnerabilities that AI-driven threat detection is beginning to address.

By Sean K.S. Shin
This blog summarizes research trends based on published paper abstracts. Specific numbers or findings may contain inaccuracies. For scholarly rigor, always consult the original papers cited in each post.

The maritime industry carries approximately 90% of global trade. A successful cyberattack on a major shipping company, a port's operational technology, or a vessel's navigation system could disrupt supply chains affecting billions of dollars in commerce. The 2017 NotPetya attack on Maerskโ€”which disrupted the world's largest container shipping company for approximately two weeks (with core IT systems restored in ~10 days) and cost an estimated $300 millionโ€”demonstrated that this risk is not hypothetical.

Yet maritime cybersecurity receives a fraction of the attention devoted to financial, healthcare, or energy sector security. Miller et al.'s systematic review is among the most comprehensive assessments of how AI can address the maritime industry's unique cybersecurity challengesโ€”challenges that arise from the intersection of legacy operational technology, remote connectivity constraints, and the physical consequences of digital compromise.

The Maritime Attack Surface

The maritime domain presents a cybersecurity profile unlike any other industry:

Legacy operational technology: Ships operate for 25-30 years. Their control systemsโ€”engine management, ballast water, steering, cargo handlingโ€”run on operational technology (OT) designed decades ago without cybersecurity considerations. These systems are increasingly connected to IT networks for efficiency and monitoring, creating pathways from internet-facing systems to safety-critical controls.

Satellite dependency: Ships at sea rely on satellite communication for navigation (GPS), weather data, and corporate communication. GPS spoofingโ€”transmitting false positioning signalsโ€”can misdirect a vessel, cause collisions, or enable piracy. Satellite communication links are bandwidth-limited and high-latency, constraining the speed and volume of security monitoring data that can be transmitted to shore-based security operations centers.

Crew digital literacy: Maritime crew members are trained in seamanship, not cybersecurity. Phishing attacks, infected USB drives, and social engineering target this knowledge gap.

Regulatory fragmentation: Maritime regulation (IMO, flag state authorities, classification societies) is fragmented across jurisdictions. Cybersecurity requirements are evolving but not yet consistently enforced, creating an uneven landscape where some vessels and ports maintain strong defenses while others remain largely unprotected.

AI for Maritime Threat Detection

Miller et al.'s review identifies three primary applications of AI in maritime cybersecurity:

Network anomaly detection: Machine learning models trained on normal vessel network traffic patterns detect deviations that may indicate intrusionโ€”unusual data flows between the bridge systems and crew personal devices, unexpected outbound communications from OT networks, or traffic patterns consistent with known attack signatures.

Predictive threat intelligence: AI models analyze global maritime threat dataโ€”reports of cyber incidents, vulnerability disclosures, threat actor behavior patternsโ€”to predict which types of attacks are most likely to target specific vessel types, routes, or ports. This predictive capability enables proactive defense measures rather than reactive incident response.

Autonomous response: For vessels at sea with limited real-time connectivity to shore-based security teams, AI-driven autonomous response systems can take immediate protective actionsโ€”isolating compromised network segments, switching to backup navigation systems, or alerting the crewโ€”without waiting for human security analysts.

Federal Cloud Security Parallels

Temiloluwa et al. examine AI-powered zero trust security for federal cloud systemsโ€”a context that shares key characteristics with maritime security: critical infrastructure status, sophisticated threat actors (nation-states), and legacy system integration challenges.

Their framework integrates three AI-driven components:

  • Continuous identity verification: AI-powered behavioral analysis that detects credential misuse even when the attacker has valid credentials
  • Automated threat intelligence: Real-time integration of threat feeds with AI analysis to prioritize alerts based on relevance to the specific deployment
  • Compliance automation: AI tools that continuously monitor system configurations against CISA (Cybersecurity and Infrastructure Security Agency) requirements, flagging deviations before they become vulnerabilities
The parallels to maritime security are instructive: both domains require security systems that operate with limited human oversight, must integrate with legacy systems, and face adversaries with resources and sophistication that exceed typical commercial threats.

Claims and Evidence

<
ClaimEvidenceVerdict
Maritime cybersecurity is a growing critical riskNotPetya precedent; increasing digitalization of vesselsโœ… Well-documented
AI improves maritime threat detection over traditional methodsMiller et al. review multiple studies showing improvementโœ… Supported
Autonomous response is necessary for vessels at seaConnectivity limitations documented; autonomous response demonstrated in simulationsโœ… Supported (rationale clear)
Current maritime cybersecurity regulations are adequateMultiple gaps identified; IMO guidelines are non-bindingโŒ Insufficient
AI-powered security is widely deployed in the maritime sectorAdoption remains early; most vessels lack AI-based securityโš ๏ธ Early stage

Open Questions

  • Air-gapped OT security: Can AI-based monitoring be deployed on isolated OT networks without introducing new attack surface? The monitoring system itself becomes a potential target.
  • GPS resilience: Beyond detecting GPS spoofing, how do we build navigation systems that are resilient to it? AI-based multi-sensor fusion (radar, inertial navigation, celestial navigation) may provide redundancy.
  • Supply chain cybersecurity: Ships interact with ports, suppliers, classification societies, and flag state authoritiesโ€”each a potential compromise vector. How do we secure the maritime supply chain without impeding the efficiency that digitalization provides?
  • Crew training at scale: The global maritime workforce exceeds 1.9 million seafarers. How do we deliver cybersecurity training at this scale across diverse cultures, languages, and educational backgrounds?
  • Insurance and liability: When a cyber incident causes a maritime casualty, who bears liability? How should maritime cyber insurance be priced when the threat landscape evolves faster than actuarial models?
  • What This Means for Your Research

    For cybersecurity researchers, the maritime domain offers a rich and under-studied application context where physical safety consequences (collision, grounding, environmental pollution) create urgency that purely digital security domains lack. The constraints of the maritime environmentโ€”limited connectivity, long system lifecycles, crew skill gapsโ€”force creative solutions that may transfer to other constrained domains.

    For maritime industry practitioners, the gap between the digitalization pace and the cybersecurity investment pace is widening. AI-driven security tools offer a path to close this gapโ€”but they require investment in network infrastructure, sensor deployment, and crew training that the industry has historically underweighted.

    For policymakers, the maritime cybersecurity gap represents a systemic risk to global trade. The IMO's guidelines on maritime cyber risk management (MSC.428(98)) provide a framework, but their non-binding nature limits enforcement. Binding cybersecurity requirements, backed by AI-enabled compliance monitoring, would accelerate the industry's security posture.

    References (2)

    [1] Miller, T., Durlik, I., Kostecka, E. et al. (2025). Artificial Intelligence in Maritime Cybersecurity: A Systematic Review of AI-Driven Threat Detection and Risk Mitigation Strategies. Electronics.
    [2] Temiloluwa, B., Ofili, Erhabor, E. (2025). Enhancing federal cloud security with AI: Zero trust, threat intelligence and CISA Compliance. World Journal of Advanced Research and Reviews.

    Explore this topic deeper

    Search 290M+ papers, detect research gaps, and find what hasn't been studied yet.

    Click to remove unwanted keywords

    Search 8 keywords โ†’