The global maritime industry is rapidly adopting AI technology to enhance operations, recruitment, and maintenance, but this swift integration poses significant cybersecurity risks. A new report reveals that up to 60% of newly disclosed software vulnerabilities related to maritime operations, both onshore and offshore, are now being weaponized within just 48 hours. This alarming trend is exacerbated by hackers increasingly employing AI to accelerate their attacks.
In 2018, the average time from the publication of new software vulnerabilities to actual attacks was 63 days. By 2024, this period had decreased to five days. Today, AI-driven tools have reduced the hacking window to under 48 hours, with systems often targeted within just 15 minutes of a flaw being identified. Tetsuji Madarame, a maritime and logistics expert and former Head of Digital Transformation and Innovation at NYK Line, emphasizes the urgency of securing AI-related assets as AI evolves from generative models to agentic and physical ones, impacting autonomous navigation and fleet operations.
Recent findings from a Cydome security research paper highlight that 87% of organizations now regard AI-related vulnerabilities as the fastest-growing risk. This underscores a troubling decline in the traditional security response timeline. While AI streamlines operations, it simultaneously enables attackers to execute “flawless deception.” Theofano Somaripa, Group CIO of dry bulk operator Newport S.A., cautions that cyber-attacks in 2026 will reflect a “shift in focus from digitalization to the radical restructuring of business models through AI.”
The report reveals that 83% of phishing emails now utilize AI to target multinational crews in their native languages, establishing trust almost instantaneously. This has led to a staggering 1,600% increase in voice phishing, or “vishing,” where AI mimics the speech patterns of C-suite executives to authorize fraudulent transactions. Katerina Raptaki, IT Manager at Greek shipping company Navios, warns that the speed of AI deployment outpaces the establishment of clear cyber accountability. In 2026, the question may not be whether the AI was at fault, but rather why it was trusted in the first place.
Data also suggests that the proliferation of edge network devices—such as routers, firewalls, and VPNs—has contributed to a decline in system trust. Cydome notes that this “digital gateway” was frequently exploited, with attacks on these devices increasing by 800% in 2025, including a direct 20% targeting of firewalls and VPNs. The report highlights a significant incident involving the hacktivist group Lab Dookhtegan, which disconnected a fleet of 116 tankers from the internet and external networks by wiping the partitions on the ships’ hard drives, leading to a total loss of connectivity, operational risks, and compliance challenges.
Øystein Brekke-Sanderud, Head of Maritime OT/ICS Security at NORMA Cyber, predicts that by 2026, the most pressing cybersecurity threats will arise from within organizational perimeters. As companies become more digitally integrated, insider risks—whether malicious, compromised, or accidental—will be increasingly challenging to detect and manage. The ability to identify subtle signals early will be crucial for resilience, rather than merely fortifying the network edge against external threats.
Panagiotis Anastasiou, Cyber Security Strategy Leader with Bureau Veritas Marine & Offshore, adds that while attacks are inevitable and becoming more sophisticated, the key differentiator will be how rapidly and effectively shipping companies can detect, respond to, and continue operations post-incident. As the maritime industry continues to embrace AI, the focus will increasingly need to shift towards creating robust cybersecurity frameworks to safeguard against the evolving landscape of cyber threats.
See also
HP Reveals AI-Driven Cyberattacks Evade Security by Prioritizing Speed Over Sophistication
Anthropic’s Claims of AI-Driven Cyberattacks Raise Industry Skepticism
Anthropic Reports AI-Driven Cyberattack Linked to Chinese Espionage
Quantum Computing Threatens Current Cryptography, Experts Seek Solutions
Anthropic’s Claude AI exploited in significant cyber-espionage operation
