Maritime cybersecurity is becoming a critical pillar of global security as ships and ports rely more on digital systems. At the crossroads of technology and policy, Strahinja “Strajo” Janjusevic is working to protect critical maritime infrastructure against emerging cyber threats. Originally from Montenegro, he studied cyber operations and computer science at the U.S. Naval Academy in Annapolis, Maryland, and is now a second-year master’s student in MIT’s Technology and Policy Program, hosted by the MIT Institute for Data, Systems, and Society. His research with MIT’s Laboratory for Information and Decision Systems and the MIT Maritime Consortium explores how artificial intelligence can improve the security of maritime cyber-physical systems, while also considering the policy frameworks needed to deploy such solutions safely.
Mitigating GPS spoofing risks for legacy vessels
Janjusevic’s thesis combines cybersecurity, AI and deep learning, and physics-based modeling to secure maritime cyber-physical systems, with a focus on large legacy ships. GPS spoofing can lure vessels off course in contested waters, creating serious national security and economic risks. His approach layers deep learning with physics-based trajectory forecasting: an internal LSTM autoencoder checks signal integrity, while a vessel dynamics model predicts movement using environmental factors such as wind and sea state. By comparing these predictions with reported GPS positions, the system can separate natural sensor noise from malicious spoofing, providing verified navigation data. The framework is designed to empower human watch standers, helping them tell apart technical glitches and strategic cyberattacks.
Linking technology, industry insight, and maritime policy
In summer 2025, Janjusevic interned with the Network Detection team at Vectra AI, where he studied risks linked to emerging technologies such as AI agents and the model context protocol (MCP), an emerging standard for AI agent communication. His work showed how MCP could be abused for autonomous hacking and advanced command-and-control operations, and it was presented in the preprint “Hiding in the AI Traffic: Abusing MCP for LLM-Powered Agentic Red Teaming.” This hands-on experience also informed the anomaly detection models used in his academic research. Alongside the technical work, he contributes to the MIT Maritime Consortium, which brings together academia, industry, and regulators to develop solutions, standards, and policy, including cooperation with international members such as Singapore and South Korea. He also helps organize the Harvard European Conference in February and the Technology and National Security Conference in April, reflecting his goal to strengthen collaboration between Europe and the United States in national security, AI, and cybersecurity.
Key takeaways:
- Maritime cybersecurity research is combining AI and policy to protect critical maritime infrastructure.
- A hybrid AI + physics approach can detect GPS spoofing on large legacy ships and support human operators.
- Industry research on AI agents and MCP highlights new cyber risks, reinforcing the need for strong standards and governance.
For more context on the same issue, here the article published by the MIT.
