If operators are penalized for taking control of autonomous vessels, could that influence critical safety decisions at sea?

That is the question raised by Asbjørn Lein Aalberg, PhD candidate in Safety Management at Industrial Economics and Technology Management, and Ole Andreas Alsos, Professor of Interaction Design at the Department of Design at NTNU and head of NTNU Shore Control Lab.

High ambitions for automation

The Norwegian Public Roads Administration is planning extensive use of automated ferry operations on the Lavik–Oppedal route from autumn. The concept involves fully autonomous ferries monitored from shore.

The contract framework includes strong incentives to minimize human intervention. Operators may face penalties of up to NOK 240 million if automation targets are not met, while all manual overrides must be logged.

From next year, at least 10 percent of departures, and five consecutive days, must be conducted without human intervention. By 2034, the requirement increases to 80 percent of departures and 20 consecutive days.

Financial penalties for intervention

These mechanisms may have unintended consequences, according to Aalberg.

“Sooner or later, situations will arise where an operator feels the need to intervene,” he says. “This may be due to system failure, lack of trust, or simply a sense that something is not right.”

Each intervention is recorded and may contribute to financial penalties.

“In practice, this creates a system where both the operator and the crew will feel the consequences if intervention happens too often,” Aalberg adds.

Alsos points to how such systems shape behaviour.

“When every intervention is logged and linked to penalties, it inevitably affects how operators think and act.”

Raising the threshold for action

While both emphasize that crews will act in clear danger, Aalberg warns about more ambiguous situations.

“In situations of uncertainty, sanctions may raise the threshold for taking control,” he explains.

He points to established findings in safety research.

“We know that accidents are particularly likely when control is transferred from automated systems to humans.”

Alsos underlines the operational implications.

“If operators do not regularly practice manual control, their skills deteriorate. Training in taking over control is essential.”

This challenge is closely linked to Aalberg’s PhD research, which examines the safety implications of bridge officers’ trust in automated and autonomous technology. His work explores how maritime operators adopt advanced systems, and whether their perceived level of trust aligns with the actual reliability of the technology.

A key premise is that safe operations depend on calibrated trust, neither over-reliance nor under-trust, but a balance that supports sound decision-making in critical situations.

Ole Andreas Alsos, professor at NTNU Department of Design and lead  Shore Control Lab. Photo: Lars Bugge Aarset/Fremtidens Industri 

Operational pressure shapes decisions

Ferry contracts already include strict requirements for punctuality and emissions—widely seen as positive for both passengers and the environment.

“Crew members are highly skilled at optimizing operations to meet these demands,” says Aalberg. “They use experience, collaboration, and judgment to reduce fuel consumption and ensure a good passenger experience.”

At the same time, these requirements influence behaviour.

“We see that contractual pressure can affect decisions—for example, whether to increase speed slightly to maintain schedules,” Aalberg notes. “This is often framed as human error, but it is better understood as a consequence of the system they operate within.”

Who is to blame when something goes wrong?

A central issue is how responsibility is assigned in complex systems involving both humans and automation.

“If a ferry fails to intervene in time and collides with the quay, who is to blame?” Aalberg asks. “Is it the technology, the operator, or the human?”

Alsos highlights a recurring pattern.

“The maritime sector has a tendency to place responsibility on the captain, even when technological or organizational factors play a role.”

NTNU Shore Control Lab. Photo: Lars Bugge Aarset/Fremtidens Industri

A flawed view of automation

Aalberg believes the current approach reflects an outdated understanding of automation.

“There is an assumption that automation can remove the human from the system,” he says. “Computers do not get tired and can execute tasks with speed and precision.”

However, he emphasizes that human competence remains essential.

“Seafarers are better at anticipating how weather, wind, and traffic affect a vessel. They manage complex situations and handle unforeseen events.”

See also: Situation awareness by design: Advancing remote operation of autonomous vessels

Encouraging intervention—not discouraging it

While the intention behind the requirements is understandable, Aalberg questions the approach.

“Financial penalties risk influencing whether operators intervene in situations where they actually should,” he says.

His conclusion is clear.

“Crew should be encouraged to intervene when necessary—not discouraged from doing so.”

About MIDAS

The MIDAS project (Humans in Future Ocean Operations) takes a systematic approach to one of the core challenges in maritime autonomy: the evolving role of people in increasingly automated systems. The project brings together industry, academia, and clusters to address human factors, trust in autonomy, and the design of effective human–machine interaction.

A key premise is the so-called “automation paradox”, that as systems become more autonomous, the human role becomes more critical, not less. By combining research, education, and industry collaboration, MIDAS aims to strengthen both safety and innovation capacity in the development and deployment of autonomous maritime operations.

See also: Teknisk Ukeblad: Griper kapteinen inn når en selvkjørende ferge feiler? (NO)