Q&A: Michelle Ramsden on AI governance in the maritime industry

Michelle Ramsden, special counsel at Jones Walker, said maritime companies adopting artificial intelligence tools should approach governance much like bridge resource management aboard a vessel: with clearly defined authority, oversight and communication protocols.

Ramsden, whose practice focuses on commercial transactions, privacy, cybersecurity and AI, spoke with The Waterways Journal about how maritime organizations can manage AI-related risks while improving operational efficiency.

Q: You recently wrote about using “bridge discipline” to govern AI. What do you mean by that?

A: I’m borrowing that concept from bridge resource management aboard vessels, where there are defined roles, communication protocols, cross-checks and escalation paths. In the maritime industry, that structure helps prevent single-point failures under pressure.

AI governance requires a similar structure. Organizations still need humans responsible for overseeing implementation, safety mechanisms and decision-making. AI tools can create problems that grow quickly if there’s no accountability structure in place.

Q: How does that translate to shoreside organizations?

A: Someone still has to be accountable. You can’t hand accountability over to technology. Organizations need designated people responsible for governance, oversight and escalation if something goes wrong.

That responsibility may be distributed among legal, IT and operational teams, but there still has to be a structure defining who handles what.

Q: You referenced four AI governance functions: govern, map, measure and manage. Can you explain those?

A: Those concepts come from the National Institute of Standards and Technology AI Risk Management Framework.

“Govern” means establishing policies, oversight structures and risk management culture around AI use. Organizations need to decide how they want to use AI and who is responsible for oversight.

“Map” means understanding where AI exists in the organization, what data it uses and what systems it connects to.

“Measure” involves testing tools for reliability, bias and operational risk. Organizations need to evaluate whether outputs are accurate and consistent with business objectives.

“Manage” is the ongoing oversight process. That includes monitoring performance, correcting issues, maintaining vendor relationships and responding to incidents when they occur.

Q: What role do contracts play when companies use third-party AI vendors?

A: Contracts are critical because AI-related risks can escalate very quickly. Companies need to understand how their data will be used, whether it could become part of a training model and what protections exist if something goes wrong.

Organizations also need to ensure liability and responsibility are clearly addressed in vendor agreements. You can’t simply rely on marketing language from a vendor’s website.

Q: Are AI contracts different from other technology contracts?

A: Yes, because AI systems can amplify errors rapidly. Small mistakes can become much larger operational or reputational problems very quickly.

Organizations need to pay close attention to intellectual property protections, sensitive information handling and liability provisions. The speed at which AI-related issues can grow makes governance especially important.

Q: Have you seen examples of AI-related risks in maritime operations?

A: There have been incidents involving falsified AIS signals, where vessels appeared to be in locations where they actually were not. That demonstrates the importance of validating information rather than blindly trusting digital systems.

AI also has the potential to amplify operational risks if organizations do not have appropriate oversight in place.

Q: Can AI also help organizations manage risk?

A: Yes. AI tools can help organizations identify anomalies, monitor output drift and detect potential issues before they become larger problems.

There are also AI testing and monitoring tools available that can help companies evaluate the performance and reliability of their systems.

Q: How are maritime companies responding to AI adoption?

A: Many organizations are interested in AI but are still trying to understand where to start. I usually encourage companies to begin with manageable use cases and build confidence over time.

Organizations do not need to solve every governance issue before beginning to experiment with AI tools. Governance frameworks can evolve as organizations gain experience.

Q: Many businesses want to know how AI can improve profitability. What are you hearing from clients?

A: AI can improve efficiency and scale operations. It can help organizations serve more clients more quickly or improve existing processes.

But companies still need people overseeing these systems. If organizations fail to govern AI properly, they may ultimately face legal or operational costs later.

Q: What advice would you give younger professionals entering the industry?

A: AI skills will likely become embedded across many professions, not just technical roles. People should develop basic familiarity with AI tools and then apply them within their specific industries or disciplines.

You still need subject-matter expertise. Someone who understands AI but does not understand maritime operations may struggle to apply those tools effectively.

Q: What is the broader takeaway for maritime companies?

A: AI is ultimately a tool, not something entirely separate from existing operational practices. Organizations can apply the same ethical and governance principles they already use elsewhere.

The companies that succeed with AI will likely be the ones with disciplined, thoughtful approaches to how those tools are integrated into operations.