1,009 Ship Officers Explain Why They Do Not Trust Autonomous Vessels

Norwegian University of Science and Technology (NTNU)

Norway is about to put autonomous ferries on the water. The first is scheduled for the Lavik-Oppedal route as early as autumn 2026, designed to operate with minimal human intervention. The Norwegian Public Roads Administration has stipulated this requirement. But the people who actually work on ships, the captains and mates who hold command on the bridge, are not convinced.

A survey of 1,009 Norwegian ship officers, the largest study of its kind globally, found broad and detailed skepticism about autonomous vessel safety. The research, conducted by Asbjorn Lein Aalberg and Professor Trond Kongsvik at the Norwegian University of Science and Technology (NTNU), identified 12 distinct categories of concern from open-ended responses. Published in the WMU Journal of Maritime Affairs, the study provides the most comprehensive picture yet of what seafarers actually think about the technology meant to replace many of their functions.

Emergency response tops the list

The most frequently raised concern was straightforward: what happens when something goes wrong? Ships encounter equipment failures, medical emergencies, fires, collisions, and weather that changes faster than forecasts predict. Officers are trained to assess these situations using judgment, experience, and direct physical interaction with their vessel and environment.

Autonomous systems are designed for normal operations. The question seafarers keep asking is whether those systems can handle the abnormal. One respondent put it concretely: in waves over four meters, experienced crews switch to a "bad weather route" based on real-time conditions, not pre-programmed parameters. An autonomous vessel following its algorithm might not make the same judgment call.

For passenger ferries, the emergency question becomes even more urgent. Evacuating hundreds of passengers requires human coordination, physical presence, and the ability to make decisions that no algorithm has been trained on because the situation has never occurred before in exactly that way.

The deskilling paradox

Perhaps the most surprising finding was how many officers worried not about the technology's shortcomings but about their own. A large proportion of respondents expressed concern about skill erosion, the possibility that as automated systems handle more tasks, the humans responsible for oversight will lose the competence needed to intervene when those systems fail.

The analogy is familiar from other domains. Pilots who rely heavily on autopilot can struggle with manual flying during emergencies. Drivers who use GPS navigation may not be able to find their way with a paper map. The maritime version is a captain who has relied on autonomous systems for routine operations and then faces a situation that demands the full range of seamanship skills that have quietly atrophied.

One respondent captured the concern plainly: "Crews are becoming lazy because they expect an alarm on absolutely everything." The issue is not that automation is bad, but that the transition period, where humans are nominally in charge but increasingly disengaged, may be the most dangerous phase of all.

Trust as a recruitment problem

The maritime industry already struggles with recruitment. Many stakeholders see automation as the solution to a labor shortage. But the study suggests this reasoning may be circular: if experienced seafarers distrust the technology, they are less likely to stay in the profession, and the industry loses the very expertise it needs to safely implement and oversee autonomous systems.

Aalberg frames this as a trust problem. "If most seafarers are sceptical about the new technology, it provides a poor foundation for recruiting competent personnel," he noted. Trust, in his analysis, is not just a psychological variable; it is a prerequisite for the human-machine collaboration that all current autonomous vessel concepts depend on. No one is proposing fully crewless ships for complex routes. The operational model is human oversight with reduced intervention. That model only works if the humans involved believe the system is trustworthy enough to rely on, and skilled enough in their own right to take over when it is not.

Parallels from medicine

Aalberg draws a comparison to healthcare, where artificial intelligence is increasingly accurate at tasks like interpreting X-rays, yet medical experts still believe they are better at assessing the whole picture, especially in complicated cases. Doctors, like ship officers, find it difficult to fully trust technology that operates in their area of expertise.

But the research evidence from medicine points in an interesting direction: outcomes are often best when humans work in interaction with AI systems, not when either operates alone. The lesson for maritime automation may be that the goal should not be minimal human intervention, as the Norwegian Public Roads Administration has stipulated, but optimized human-machine collaboration.

What the study does not answer

The survey captures attitudes, not outcomes. It tells us what seafarers believe about autonomous ships, not whether those beliefs are correct. It is possible that autonomous systems will, once deployed, prove safer than human-operated vessels in many scenarios. The data to answer that question does not yet exist because large-scale autonomous vessel operations have not begun.

The sample is exclusively Norwegian. Norway is a global leader in maritime autonomy research and deployment, so its seafarers may be more informed about the technology than their counterparts elsewhere, but their concerns and cultural context may not generalize to other maritime nations.

The study also does not capture passenger perspectives or those of other maritime stakeholders, such as port authorities, insurers, or regulatory bodies, whose trust will also be essential for the technology's adoption.

What the study does establish, convincingly, is that the people with the most direct experience operating ships have specific, articulated, and widespread concerns about autonomy. Whether those concerns are addressed through better technology, better communication, or better human-centered design will determine whether autonomous shipping succeeds not just technically, but operationally.