625

1 INTRODUCTION

The International Maritime Organization (IMO)

Maritime Autonomous Surface Ship (MASS) Joint

Maritime Safety Committee (MSC) - Legal Committee

(LEG) - Facilitation Committee (FAL) Working Group

on the 109th session stated within the principles of the

MASS Code draft that “There should be a human

master responsible for a MASS, regardless of mode of

operation; regardless of mode of operation, the master

of a MASS should have the means to intervene when

necessary” [1]. These principles should continue

unchanged as per current IMO's vision of adopting a

non-mandatory MASS Code by 2026 and a mandatory

MASS Code in 2032 [1]. This means that for a fully

autonomous ship, a human master will remain in

command, ultimately responsible for the safety of the

MASS vessel. This person will delegate and share

responsibility with an Autonomous Navigation

System (ANS) for the conduct of a MASS vessel.

Therefore, even for fully autonomous ships, humans

will still be mandated to be in the loop, operating in

monitoring or supervising mode. In addition,

individual masters may oversee one or more MASS

vessels at any given time, remotely viewing or

controlling vessels from a Remote Operations Centre

(ROC) to “intervene when necessary” in case the ANS

makes an incorrect decision [2, 3].

The above demonstrates that regulations,

standards, competences and certification for Remote

Operator (RO) and ROC personnel will be as crucial as

technical regulations and standards developed for

MASS [4, 5]. The share of responsibilities during the

hand-over between a RO to an ANS for the conduct of

the ship and vice versa will require Human-Autonomy

Teaming (HAT) to perform Uncrewed Autonomous

and Remote Surface Operations (UARSVO) [6]. HAT is

Developing Human-Autonomy Teaming Strategies

for Maritime Cyber Security Resilience in Uncrewed

Autonomous and Remote Surface Vessel Operations

J.D. Palbar Misas, K. Tam & K. Jones

University of Plymouth, Plymouth, United Kingdom

ABSTRACT: The development of new technologies and digital capabilities for Uncrewed Autonomous and

Remote Surface Vessel Operations (UARSVO) is driven by various industry stakeholders. This evolution impacts

the maritime industry's human role, transforming from Human-Autonomy Hybrid (HAH) to Human-Autonomy

Teaming (HAT). Human-Autonomy Collaboration (HAC) is vital for maritime safety, security, and sustainability,

particularly in light of increasing cyber incidents in remote operations, which necessitates greater cyber resilience

due to technology at sea and ashore. This paper aims to provide a holistic socio-technical approach to investigate

and present an overview of the current state-of-the-art research, focusing on the human perspective in maritime

cyber resilience for UARSVO. The view is developed using 76 semi-structured interviews with participants from

various stakeholder groups across the Maritime Autonomous Surface Vessel (MASS) industry worldwide.

Findings provide unique, holistic, and human-centred operational strategies for maritime cyber resilience, as well

as the perceived training required to enhance these operations.

http://www.transnav.eu

the International Journal

on Marine Navigation

and Safety of Sea Transportation

Volume 19

Number 2

June 2025

DOI: 10.12716/1001.19.02.35

626

the interdependence of human(s) and autonomous

agent(s) working collaboratively but, with separate

roles, to attain a common goal [7-9]. Thus, clear roles

and responsibilities for all agents must be established

to attain seamless, safe, secure and sustainable

operations for future MASS vessels, especially when

engaged in international commercial voyages [10].

Due to the increase in connectivity and complicated

software, no part of the industry is, nor will be,

immune to cyber threats. As demonstrated with real-

world maritime autonomous systems in [11], multiple

attacks can be performed utilising Adversarial

Artificial Intelligence (AAI) to confuse MASS- running

AI without HAT. When considering more general

cyber threats in the maritime industry, from October

2023 to October 2024 31% of maritime professionals

reported a cyber-attack, indicating a rise of 14% over

the previous five years [12]. As completely cyber-

secure maritime systems are not practically feasible, a

key success for future operations is cyber resilience.

Providing resilience is key to enabling autonomous

systems and operators to reduce and minimise cyber-

attacks but also continue operating safely during an

active attack. Cyber resilience can be enhanced by

developing proactive strategies including response

procedures during cyber incidents. While it is difficult

to train for this during a real incident, with the use of

response plan scenarios played out (for example in a

simulator), training cyber resilience becomes more

feasible [13].

To achieve a maritime cyber resilient workforce,

stakeholders and Maritime Education and Training

Institutes (METIs) should collaborate and develop the

appropriate training requirements for UARSVO. This

would expand and introduce new and essential social-

technical skills and behavioural markers [14, 15]. As

this training is meant for people, using methods like

the Human Centred Design (HCD) process when

developing training can increase understanding of the

operational needs of ROC operators (i.e. end-users).

Moreover, to achieve a balance between the innovation

technology developed and operational safety, ROC’s

top management level will need to serve as a link for

the UARSVO stakeholder ecosystem so that

technology developers (software and hardware

manufacturers) can receive ROs (end-user) feedback

on products before and after implementing technology

[16].

UARSVO requires interdisciplinary collaboration to

integrate technical and operational skills within a

cooperative feedback loop to foster a shared vision of

operations across organisations and enhance the

resilience of proactive organisations [17]. As an

example of a holistic solution, [18] uses HCD combined

with the Desing Research Methodology (DRM), as a

business and technical approach which includes the

user need, to develop and conduct maritime cyber

resilience training for crewed operated vessels with the

participation of several maritime stakeholders’ and

operators. That said, that previous study focuses on

crewed operations, differs significantly to this study,

which focuses purely on uncrewed.

This study explores which training strategies could

best enhance HAT in maritime cyber security resilience

for UARSVO. To achieve this, 76 semi-structured

interviews were held with participants from multiple

stakeholder groups across the MASS industry,

including international organisations, research and

academia, national organisations, industrialists and

resource providers, shipping companies, and members

of the future workforce. This large, multi-disciplinary

engagement provided a never-before-attempted

holistic socio-technical view towards the future

development of a HAT training framework in

maritime cyber security resilience for UARSVO. The

remainder of this paper follows the following

structure: Section 1 introduces the context and draws

the research objectives. Section 2 provides a literature

review of HAT for UARSVO, HAT for maritime cyber

security resilience and HAT maritime cyber security

resilience training in UARSVO. Section 3 states the

qualitative methodology and methods used for

research as well as ethical considerations. Section 4

presents the results obtained per stakeholder group,

and the findings are discussed in Section 5.

Section 6 outlines the limitations of the research and

future work. Section 7 summarised key findings for the

conclusion.

2 BACKGROUND

2.1 HAT for UARSVO Overview

The autonomy of a maritime vessel is classified by the

degree it makes decisions and executes actions by

itself. It must also, critically, perform these actions at

the appropriate time without human intervention [19,

20]. As defined by [21], autonomy is considered a

process as both autonomy and automation are

somewhat overlapping terms that some use

interchangeably. In this paper, if a machine or a

computer performs operations without human control,

this is referred to as automation. This means that

sophisticated automation could overlap with the term

autonomy [21, 22]. Although people seem to perceive

“autonomy” as taking humans out of the loop,

sometimes implying technology is superior to humans,

the irony of automation is that if technology fails,

humans will always be needed to take over [23].

Additionally, if technology acts unexpectedly for the

user, causing a dangerous situation, this could be

named as automation surprise [24]. The incident report

findings provided by the National Transportation

Safety Board (NTSB) in 2018 demonstrate the concept

of automation surprise as there was a perceived loss of

steering by the crew and subsequent mishandling of

steering controls during the transfer of control. This

caused the USS John S McCain in the Singapore Strait

to veer off course and consequently collide with the

tanker Alnic MC. This incident, which tragically led to

the loss of life of seafarers aboard the USS John S

McCain, highlighted the need for improved training

when systems are perceived to malfunction [25].

However, a malfunction and a cyber attack are

different. This paper builds on this with the hope that

research in cyber resilience training can help prevent

tragic incidents involving autonomy and cyber- attacks

instead of being implemented after a tragedy.

From a socio-technical approach, and as seen with

the incident case above, complex systems must be

designed so that the technical and human elements can

work in unison to achieve safe and effective operations

627

[26]. This means that systems are designed to enhance

the adaptability human performance brings to

dynamic and critical environments when oversight is

required [27]. For this, there are three paradigms to the

overall Human-Autonomy Relationship (HAR):

Human-Autonomy Collaboration (HAC), Human-

Autonomy Teaming (HAT) and Human-Autonomy

Hybrid (HAH). At a strategic level, HAC encompasses

any collaborative cooperation between humans and

autonomy to achieve a common goal; at a tactical level,

HAT significates that there is partial human authority

as human and autonomy work as a unit with distinct

roles each; at an operational level HAH autonomy

works complementary and interdependently as an

integrated unit with the human having the full

authority in decision making and autonomy for

decision-support [28, 29]. In the maritime industry for

MASS vessels, the amount of control the systems have

within the Operational Design Domain (ODD) and

established performance capabilities and limits within

the Operational Envelop (OE) is determined by the

IMO’s degree of autonomy (as shown in Table 1) which

sets the boundaries for each agent [30]. From the

human perspective, the degree of autonomy helps

define ROC operators’ competences and certification

needed for UARSVO [4, 5, 31].

Table 1. Compilation of different Degrees of Autonomy

(DOA) based on [32] and [33].

DOA

IMO

Bureau

Veritas

DNV

ONE SEA

SARUMS

Lloyd’s

D-0

N/A

System

operated by a

human.

Manually

operated

function.

Basic

operation.

Crewed

Manual

steering.

D-1

Ship with

automated

processes

and

decision

support.

Systems

directed by a

human.

System

decision

supported

function.

Assisted

operation.

Operated

On-

board

decision

support.

D-2

Remotely

controlled

ship with

seafarers

on board.

Systems

delegated by

a human.

System

decision

supported

function with

conditional

system

execution

capabilities.

Partial

automatio

Directed

On and

off-

board

decision

support.

D-3

Remotely

controlled

ship

without

seafarers

onboard.

Systems

supervised

by a human.

Self-

controlled

function with

human in the

loop.

Conditiona

automatio

Delegated

“Active”

human

in the

loop.

D-4

Fully

autonomo

us ship.

Fully

automated

systems that

only require

human

intervention

in case of

emergency.

Autonomous

function

without

human

intervention.

High

automatio

Monitored

Human

in the

loop.

D-5

Autonomo

us.

Autonomo

Autono

mous.

D-6

Fully

Autono

mous.

Frameworks by [34], [35], [36], [37], [33], [38].

2.2 HAT Maritime Cyber Security Resilience for

UARSVO

Between the loss scenarios and causal factors identified

in autonomous ships on the EU H2020 MOSES project,

43% of the causal factors of a loss scenario were

attributed to cyber-attacks or connectivity issues [39].

This demonstrated that cyber resilience for UARSVO is

a major concern and a crucial area of improvements as

a continuous iterative process. In a survey conducted

by DNV with maritime expert professionals, 95% of

roughly 500 participants, stated that further

collaboration is needed within the maritime industry to

protect critical infrastructure from a cyber incident

[40].

Within the current maritime international

standards, the Resolution MSC.428(98) – Maritime

Cyber Risk Management in Safety Management

Systems has been adopted by the IMO’s MSC in 2017

and entered into force in 2021. This resolution

mandates that currently operating ships must assess

cyber risks in their company’s Safety Management

System (SMS) as per the International Safety

Management (ISM) Code [41]. To further provide

guidance for maritime cyber resilience operations, the

IMO in 2022 released the Guidelines on maritime cyber

risk management under the MSC-FAL.1- Circ.3-Rev2.

The document's intent is to embed cyber resilience

principles for operations as states a high-level

perspective of maritime cyber risk management and

adopts the National Institute of Standards and

Technology (NIST) framework with Identify, Protect,

Detect, Respond and Recover as the elements of

maritime cyber management [42]. Both technological

resources (e.g. hardware and software) and workforce

resources (e.g. ROC operators) are considered crucial

for organisations' ability to achieve cyber resilience [43,

44].

From the workforce resource perspective, achieving

UARSVO-related competences (Knowledge, Skills,

Abilities and Behaviours) for maritime cyber security

will be required for ROs as well as for UARSVO

companies to maintain cyber security standards [33,

45]. At the international level, RO competences can be

seen within the 2022 standard DNV-ST-0324

competence of ROC operators which introduces cyber

security and cyber-attack competences under the

emergency handling section [5]. Further clarification

of competence standards for MASS RO was provided

by the MASSPeople working group's submitted

document for the IMO Intersessional Working Group

on MASS in September 2023. This document includes

cybersecurity competences at the (1) support level for

implementing cyber-protection measures, (2)

operational level for operating cyber-protection

measures and responding to abnormal and emergency

scenarios (including a cyber-attack), and (3)

management level for managing cyber-protection

measures [46]. Academically, [47] demonstrated a

potential solution to develop standardised cyber

security competences with the use of the NIST cyber

security framework, which works with IMO’s interest

in NIST. However, for those RO competences to be

developed, effective training must be implemented

with the use of appropriate methods. As highlighted in

the recommended practice by DNV [4], periodical

training is also an essential part of continuous

development as well as ROs competence retention. As

628

an example, the training course program created by

[48] with the use of a survey provides a proposal on

how MASS could be integrated within METIs and

highlights the identification, prevention and mitigation

of cyber-attacks within the main four competences.

Moreover, their course guide proposal indicated in the

course timetable proposal that 35% of the total hours

were specified to be dedicated to cyber security and

data management. However, that training course

programme was created only using a questionnaire,

and there was no participation from maritime cyber

security experts or international maritime Subject

Matter Experts (SMEs). This paper fills in the research

gap for the development of more thorough maritime

cybersecurity competences that could serve as a

foundation for constructing a training framework

focused on maritime cybersecurity resilience for RO.

2.3 Training for Maritime Cyber Security Resilience

UARSVO

In 2015, a collision in the River Humber (UK) between

a pure car carrier and a RoPax ferry occurred as the

pilot experienced disorientation due to “relative

motion illusion” as well as visual constraints (caused

by the limitations of bridge design), which led him to

take incorrect actions [49]. This accident serves as an

example of how spatial disorientation can affect

navigators in crewed and remotely operated vessels

perceiving motion in a direction that is perpendicular

to the actual true motion. This could also cause

automation surprise if the RO perceives the

instrumentation performing unexpectedly when

compared to their own visual cues [50]. As shown by

the accident and examples in [51], MASS can be highly

susceptible to cyber-attacks and that the complexity of

operations could increase if a cyber incident (as defined

in Table 2) was the causal factor. An example of this can

be seen with the experienced mariners’ response

towards different cyber incident scenarios onboard

traditional ships replicated on a full bridge ship

simulator using different types of cyber-attacks. Some

public scenarios of cyber-attacks include Global

Navigation Satellite Systems (GNSS) spoofing and

jamming, jamming of rudder and engine controls,

ransomware attack, taking control of the Ballast Water

Management System (BWMS) and Automatic

Identification System (AIS) spoofing [13, 52-56].

Currently, for the cyber incidents shown above,

there have been some cyber incident management

response plan procedures developed for crewed

commercial vessels, such as detailed best practice

checklists for detecting and responding to a cyber

incident by [57] and the Cyber Emergency Response

Procedure (CERP) by [58]. These provide different

steps to follow with the use of a flow chart.

Additionally, the CERP framework demonstrates that

there are three clear types of iterations that can

followed in the face of a cyber incident, and each one

requires a different maritime stakeholder involved to

either recover the vessel with normal operation or

maintain the ship’s operation in a reduced mode [58].

The Maritime Cyber Security (Marcy) training

program involved different maritime stakeholders for

crewed vessels has been developed and evaluated [59,

60]. The Cyber-SHIP lab also provides different types

of multi-disciplinary maritime cyber security training

for a wide range of participants [61]. However, while

current cyber resilience management and cyber

resilience training have been developed for crewed

vessels, there is a unique gap to be addressed when

mariners are transferred to operate vessels remotely for

UARSVO. This is due to a reduction in Situational

Awareness (SA) as ROs are solely reliant on data,

technology, and connectivity when operating from an

ROC. Moreover, a further reduction in SA that ROs can

occur when facing a cyber incident as there may be no

physical crew onboard to witness or respond to these

incidents directly (e.g. for manual control). This paper

attempts to address this need for existing mariner

training on cyber security to be expanded and adapted

to ensure ROs are equipped with the required

competences to interact with digital systems for

effective HAT in maritime cyber security resilience for

UARSVO [13].

Table 2. Definitions for cyber-related terms within MASS are

based on [3].

Term

Definition

Cyber

Security

The security of digital systems, software, and the

data/information they hold across the autonomous ships

themselves and related infrastructure such as Remote

Control Centers (ROCs).

Cyber

Resilience

Resilience of MASS and ROC systems ensures that key

operations can safely continue despite an active cyber

attack affecting systems and/or data.

Cyber

Defense

Digital defenses for MASS and ROC are solutions that

prevent a cyber attack from happening, entirely or

partially, such as firewalls.

Cyber

Physical

For the robotics and operational technology (OT) aspect

of MASS, cyber physical refers to the risk of physical

impacts (e.g., delay, physical damage, loss of cargo) due

to a cyber- related incident.

Cyber

Attack

A cyber attack for MASS in this chapter will by default

imply cyber physical attacks. If an attack is Information

Technology (IT) only, it will be stated as an IT cyber

attack. This implies an attacker and excludes error.

Cyber

Incident

An active cyber attack, passive attack, or error involving

digital assets (e.g. the latency of messages, and the contest

of messages or connectivity availability between ROC

and the MASS vessel)

3 METHODOLOGY

For this research, an inductive approach (or “bottom-

up”) qualitative methodology was used as it allows an

analyst to identify patterns of why a phenomenon is

happening through data [62-64]. This approach

provided flexibility, developing themes within the

data. It also added credibility and validity to the

research outcomes without using pre-existing coding

framework constraints [65-67]. With this approach, this

paper examines which strategies could best enhance

HAT in maritime cyber security resilience for UARSVO

from both an end-user and an interdisciplinary

perspective. This contributes to the proposed hybrid

Design Research Methodology (DRM) and Human

Centre Design (HCD) (International Organisation for

Standardization (ISO) standard 9241-210:2019)

iterative approach for the development of a HAT

training framework in maritime cyber security

resilience for UARSVO [68, 69]. DRM performs as the

primary expansive methodological framework for

design research in this research, which can be applied

to a variety of research disciplines. DRM methodology

provides an effective rigorous approach and an

629

efficient measurable criterion throughout the different

stages of this methodology to answer and assess the

development of each research question [68]. However,

as DRM is not end-user focused, HCD methodology

has been integrated to embed the user in the design,

development, testing and evaluation iterations

between activities so that the user, the user tasks, and

the environment can be explicitly understood. The

operator’s involvement in activities with all

stakeholders (within UARSVO) through all research

stages will help provide the whole user experience, as

well as multidisciplinary perspectives and skills [69].

3.1 Selection and Sampling of Interview Participants

The selection of interview participants for this research

used purposive sampling, a commonly used

qualitative research analysis for selecting a specific

sample of participants from the overall population [70-

72]. This method helped attain rich and in-depth data

on this paper’s research topic from 76 representatives

across key stakeholders. All participants selected were

Subject Matter Experts (SMEs) from different maritime

organisations across the MASS industry, as shown in

Table 3 [73]. Although all participants currently work

in the maritime (except one in the space industry),

some participants had previous experience in the

aviation (n=3), space (n=2) and automotive (n=2)

industries and maritime cyber security (n=18). This can

be seen in Table 4, within participants' varied

experiences, 21 participants were former mariners

holding a Certificate of Competency (COC) as either a

Deck (17 participants), Engine (3 participants) or

Electronic Technical (1 participant) Officer occupying a

key position/role within different maritime

organisations. Participants’ experiences as a group

(experience participants had accumulated within

maritime and/or other industries) varied, as shown in

Table 4, with 43% of them having more than 15 years

of experience as a group. Moreover, as shown in Figure

1, the participants' occupational location was from 16

different nations across Europe, Asia, North America

and Oceania.

Given the diversity of specialised groups that

needed to be included, it was not possible to do this

research with less than ten interviews from only one

key MASS stakeholder. To best capture opinions across

many topics and from many stakeholders, the

snowball sampling method was used to gain a novel,

holistic, view [74]. However, unlike previous research,

which snowballed for one topic, this was an iterative

process where several topics were snowballed.

Interviews continued until data saturation was

achieved, with fewer new themes or data emerging.

This showed that homogeneity amongst participant

data, as well as consistency in findings from the

different stakeholder groups, could be achieved [75,

76]. It was found that, for the umbrella topic of ROC

cyber security training, 76 interviews were needed to

fully explore and identify the diverse perspectives and

shared themes across multiple international and

interdisciplinary stakeholder groups (as well as unique

group- specific insight) to obtain data saturation [77].

Table 3. Interview participant’s demographics (source:

author).

Key Group

of MASS

Stakeholders

Response N

per Group

MASS Stakeholders

Response

International

Organisations

Non-Governmental

Organisations (NGO)

Classification Societies

Insurance

Research &

Academia

Maritime Cyber Security

Unmanned Aerial Vehicles

Maritime IoT and Mobile

Networks

MASS

Maritime Business and Supply

Chain

Maritime Regulation and Policy

Marine Biology

National and

Regional

Government

Coastal Administration

Accident Investigation

Hydrographic Office

Military Navy

Port Authorities

Pilotage Service

Industrialists

and Resource

Providers

Technology Provider

Ship Builders

Maritime Cyber Security

Service

CoC Training Provider

RO MASS Training Provider

Shipping

Companies

and Future

Workforce

Shipping Companies

MASS Shipping Companies

Deck Cadets

Total N of Participants

Table 4. Interview Participant’s demographics (source:

author).

Demographics

Variable

Response

Proportion %

Merchant Navy

Certificate of

Competency

(CoC) holders

and Candidates

Deck Officer or above

22.37

Engineer Officer or above

3.95

Electronic Technical

Officer

1.32

Deck Officer Cadets

15.78

None CoC Holders or

Candidates

56.58

Experiences as

Group

Represented (in

Years)

0-1 year

13.16

2-5 years

10.53

6-10 years

9.21

11-15 years

23.68

Above 15 years

43.42

Figure 1. Countries covered according to participants

occupational location (source: author).

3.2 Data Collection and Ethics

Semi-structured interviews, also known as moderately

structured interviews were our primary means of data

collection [78]. This fostered an opportunity to address

630

questions from a clear list of thematic issues with

flexibility in exploring new issues as they arose

dynamically [79, 80]. It gave the interviewees an

opportunity to develop ideas and provide open-ended

answers so that participants could elaborate on points

of interest through personal experiences [63, 81]. To

ensure that during the investigation, all perspectives

were sufficiently covered, an interview guide was used

(see topics in Table 5) [81]. This has been developed

using previous research by [13] on training for ROs in

maritime cyber resilience for UARSVO so that the

method used was in line with the DRM and HCD

methodologies' iterative processes [68, 69]. During the

interview, a Microsoft PowerPoint was shared to

introduce IMO’s vision for the development of the

MASS Code, responsibilities for MASS, and degree

levels of autonomy as per IMO and interview guide

questions. This ensured a common base of

knowledge for all participants before starting.

Prior to starting each semi-structured interview,

participants obtained via email in English, the

established working language for this study:

1. A participant information sheet (with research aims

and objectives; contact details of the research

investigator and director of studies), semi-

structured interview questions guide (for

participation involvement);

2. Consent form (with rights to withdrawal,

confidentiality and anonymisation) and verbal

confirmation were asked to acknowledge that all

documents were understood as well as to clarify

any concerns or questions [65, 82].

With consent, data in the form of audio and video

recordings were collected for each interview.

Recordings were automatically transcribed via

Microsoft Teams, used for both online (57) and in-

person (19) interviews. All these interviews occurred

between April 2024 and October 2024. The interviews

lasted from 26 minutes to 2 hours and 50 minutes, with

a mean of 54 minutes. The total time accumulated

performing all interviews was 68 hours and 40 minutes

(2.86 days).

Due to confidentiality, interviewees' responses

were not given by any identifying details or

pseudonyms. To prevent any identification, response

comments were provided as a group and gender

neutral for General Data Protection Rights (GDPR) to

make results public [83].

Table 5. Semi-Structures Interview Guide Questions.

Theme

Questions

1. Opening

Questions

1.1 Before we start, do you have any questions

regarding the consent form and/or information sheet?

Or any questions or comments regarding this

research?

1.2 When and why did you begin your career in the

Maritime Industry? What made you start a career

pursue a career in the maritime industry?

1.3 What is your current Expertise? How is a normal

day at your current job?

1.4 Experience (years/months)? Type of ships

(age/flag) or companies you have worked on?

2. HAT

Relationship

for UARSVO

2.1 What do you think the human role and

responsibilities will be for navigation officers in

Autonomous and Remote Vessel Operations? Which

parts will remain the same, or be different, and why?

2.2 What challenges do you foresee for operators of

remote-controlled vessels (e.g. trust and reliance on

navigation equipment information from sensors to

safely navigate)? What type of standards and training

do you consider vital for safe navigation when

operating from a Remote Operations Centre?

2.3 What is your view on the Human-Autonomy

Teaming relationship for Autonomous and Remote

Surface Vessel Operations (e.g. decision-making

process for different modes of human operation such

as monitoring, supervision and remote control)? Are

there particular operations that will be affected more?

If so which ones and why?

3. HAT for

Maritime

Cyber

Security

Resilience

3.1 What types of cyber threats could an autonomous

vessel be vulnerable to (e.g. GPS Spoofing)? What

catastrophic events could this cause?

3.2 How could this attack be detected, reported,

contained or eradicated by a Navigator at a Remote

Operations Centre?

3.3 How do you think Human-Autonomy Teaming

can enhance maritime cyber security resilience (e.g.

strategies to resolve a potential machine error and/or

reduce the magnitude of the incident)?

3.4 What type of training and procedures could help

the operator to enhance maritime cyber security

resilience? Would this training involve personnel

across all levels of the organisation with external

organisations (e.g. during cyber incident

management)?

4. Debrief

and Future

Research

Engagement

4.1 Thank you for your participation. Do you have

any questions or comments regarding this research?

4.2 Would it be possible please to keep your name

and contact details to be contacted to participate in

the next stage of this research?

3.3 Data Analysis and Fieldwork

The data from the semi-structured interviews was

examined with thematic analysis, as per [75] ‘s six- step

process. This allowed us to identify and organise

patterns in content and meaning [84, 85].

Familiarisation of data began by checking the

transcribed transcripts saved as unique documents

alongside the audio recordings. These transcripts were

made available to participants for review, and the

process was further supported by written field notes to

ensure the quality and reliability of transcriptions [86].

Secondly, the anonymised and reviewed transcriptions

were uploaded to NVivo software for the Generation

of Codes, Combining Codes Into Themes, Reviewing

Themes, Determine Significance of Themes and

Reporting of Findings [112]. The gathered data was

analysed descriptively (focused on what the data

shows) and interpreted inductively (focused on why

the patterns exist) [66, 75, 87]. On completion of the

data analysis, four main themes arose in conjunction

with sub- themes as presented in the contextual model

in Section 4 below, which serves as the foundation for

presenting the results obtained.

4 RESULTS AND ANALYSIS

The interview themes that emerged when discussing

HAT strategies for maritime cyber security resilience in

UARSVO, as shown in Figure 2, were used to divide

this research into four topics: “UARSVO” overview,

“Maritime Legislation for UARSVO”, “Maritime Cyber

Security Resilience in UARSVO” and “Maritime Cyber

Security Training for UARSVO”. Additionally, a graph

demonstrating the occurrence difference between key

groups of MASS stakeholders is shown for each theme

(as per Figure 3) and for sub-themes within each theme

631

section (as per Figures 4, 6, 8 and 9). However, the

“other industries” group was omitted from the graphs

due to a low number of participants. Results from this

group were embedded as occurrences within the table

of total occurrence for each theme and sub-theme

(which is below the graph) in each Figure. As shown in

Figure 3, of the main four themes within the table of

total occurrences, “UARSVO overview” was discussed

the most, with 1,016 occurrences; on the other hand,

“Maritime Cyber Security Training for UARSVO” was

discussed the least, with 411 occurrences. Furthermore,

amongst key groups of MASS stakeholders, “Maritime

Legislation for UARSVO” was discussed the highest by

national and regional organisations, with 247

occurrences, whereas “Maritime Cyber Security

Training for UARSVO” was discussed the least by

international organisations, with 41 occurrences.

Figure 2. Emerged main themes and sub-themes with the use

of NVivo from performed semi- structured interviews

(Source: Author).

Figure 3. Main themes identified with the use of NVivo from

semi-structured interviews performed (Source: Author).

4.1 Uncrewed Autonomous and Remote Surface Vessel

Operations

Figure 4 demonstrates that of the four sub-themes

within “UARSVO”, “HAC for UARSVO” was

discussed the most, with 408 occurrences. Conversely,

“Other industries’ adoption” was discussed the least,

with 103 occurrences. Furthermore, amongst key

groups of MASS stakeholders, “HAC for UARSVO”

was discussed the most by industrialists and resource

providers, with 86 occurrences, whereas “Other

industries’ adoption” was discussed the least by

shipping companies and future workforce, with 2

occurrences.

Figure 4. Main sub-themes emerged from the theme

Uncrewed Autonomous and Remote Surface Vessel

Operations (Source: Author).

4.1.1 Challenges for UARSVO

56 participants (73.69%) were mainly concerned

about operational challenges for MASS, especially for

coastal waters and port arrival/departure operations

such as pilotage operations. The primary reasons were

concerns about MASS performing safe passage

through heavy traffic in different points around the

globe, while being in compliance with the Convention

on the International Regulations for Preventing

Collisions at Sea 1972 (COLREGS). In addition, there

were concerns between crewed and uncrewed vessels

in conjunction with complex vessel manoeuvres such

as leaving anchorage (e.g., Singapore Strait), narrow

channels for vessels in head-on interactions, mooring

operations, and adverse weather conditions [88].

Participants expressed concerns about

communications with the MASS vessel from ROC and

the MASS vessel communicating with other vessels,

Vessel Traffic Service (VTS) and other entities.

Additionally, significant challenges for the existence of

uncrewed vessels included the vessels' constant

maintenance requirements during international

voyages.

Human factors and lack of HCD when developing

systems were ranked second among challenges as

operators may experience a loss of realism due to

digitisation of operations, physical detachment, and

loss of situational awareness. Participants also

discussed MASS technology readiness challenges in

sensor discrepancy for data received or lost, as well as

latency or connectivity availability of the MASS vessel.

As per participants' views, the challenges of

technology readiness extended to trust and reliance on

automation and navigation equipment and

information from sensors. Additionally, the lack of

workforce readiness (future workforce shortage and

unqualified workforce) and infrastructure supporting

these operations were agreed upon as the maritime

industry is slowly keeping up with technology.

Moreover, participants perceived that the maritime

industry is not yet mature in cyber security,

operational law and contractual law, MASS technology

readiness costs (including economic consequences due

to loss of MASS), and the industry stakeholder shift

(such as the product and service provider becoming the

new ship owner of MASS).

632

4.1.2 Opportunities for UARSVO

Participants spoke of major opportunities that

could be available with the operational application of

UARSVO, highlighting that these opportunities will

depend on vessel type, operation, location, weather

and traffic. Individuals indicated that, based on the

relevant experience, opportunities have risen for

smaller ships in size, such as ferries, survey vessels and

cargo vessels operating in or near coastal waters with

better connectivity, lower maintenance required, and

more redundancies (to physically intervene whether it

is to repair or hot swap the vessel). It was also

highlighted that, globally, zones for UARSVO could

include coastal waters and open seas, as these were

seen as less risky environments, especially for the parts

of the voyage where the autopilot onboard a vessel is

already being used. The second most referenced topic

was the opportunity to make safer and more efficient

operations. Many participants discussed how

autonomy could reduce human error with situational

awareness optimisation, vessel predictive

maintenance, reduction in vessel operational costs and

crew costs, as well as more sustainable operations due

to vessel optimisation. An additional suggestion was

opportunities to create dedicated shipping lanes

and/or vessel lights and shapes (as per COLREGS) for

MASS vessels when operating remotely or

autonomously for safer operations in a mixed

environment with crewed vessels.

4.1.3 Human-Autonomy Collaboration (HAC) for

UARSVO

In this sub-theme, the most frequently referenced

topic was the significance of the HAT symbiotic

relationship. Participants stated that the ANS would

carry out 99% of the tasks according to the goals and

parameters set by humans, with just 1% reserved for

critical decision-making by humans in situations

where the ANS is too limited to make a safe decision.

Individuals stated that this 1% was as valuable as the

99%, as the human in the loop could prevent large

catastrophic scenarios with a single intervention. For

this, it was stated that gradual trust in technology

increases as technology readiness and maturity levels

are proven with performed safety and security

reliability records as well as communication

transparency of decision-making from ANS to ROs at

ROC and vice versa.

From the participants’ perspective, this maturity

level could help to attain clarification in liabilities for

insurance coverage when ANS is with the conduct of

the vessel and an accident occurs due to automation

surprise. Additionally, individuals stated that clear

alarms and indications from the ANS to the RO for

seamless handovers when transitioning between

degrees of autonomy could help overcome these

challenges.

As shown in Figure 5 within HAC, according to

IMO degrees of autonomy (as shown in Table 1),

participants expressed that at degree 0 should be

added as there is no autonomy in direct control (e.g.

manual control). Participants stated that degrees 1 to 3

should be classified as HAH as the navigator has the

conduct of the ship and could have ANS for decision

support. In degree 4, it was stated that it should be

classified as HAT due to ANS becoming similar to a

junior Officer of the Watch (OOW) with the conduct of

the vessel with oversight by the Master and RO. The

RO oversight was stated as standby (availability at

ROC), monitoring (assessing ANS) and supervision

(verification of ANS and assisting with decision-

making). Additionally, per IMO’s definition for

degrees 3 and 4, from participants' perspectives, this

was defined as uncrewed rather than unmanned, as a

passenger ferry could operate autonomously without

crew onboard but transporting passengers.

Figure 5. Maritime Human-Autonomy Collaboration (HAC)

framework according to the IMO degrees of Autonomy

(Source: Author).

4.1.4 Other Industries’ Adoption

Aviation was the industry most frequently

mentioned among other industries' adoption,

regarding air traffic control operations. In particular,

individuals viewed their highly standardised

communication, handover procedures implemented,

and duration of shift rotations according to human

cognitive ability as valuable practices to adopt within

UARSVO for RO training. This extended towards

aircraft pilots' training and refreshment training using

scenarios and simulators. Second-ranked was the

space industry, where some participants with relevant

backgrounds stated that humans always remain in the

loop for space operations. In regard to space training,

the just-in-time approach to training adopted by the

National Aeronautics and Space Administration

(NASA) was suggested for UARSVO as it is focused on

providing the specific training, information and

resources when needed. Furthermore, from

participant's perspective, RO training conducted for

military operations for Unmanned Aerial Vehicles

(UAV) offers approaches that could be effective for RO

in UARSVO, such as the adopted specialised UAV

operator training approach to acquire the required

airmanship skills as UAV operators without previous

experience as aircraft pilots.

4.2 Maritime Legislation for UARSVO

Figure 6 illustrates the four sub-themes discovered in

this research within the theme “Maritime Legislation

for UARVO”. Of these four, “Governing Bodies” was

discussed the most, with 474 occurrences; on the other

hand, “Remote Operator Role and Responsibilities”

was discussed the least, with 242 occurrences.

Furthermore, amongst key groups of MASS

stakeholders, “Governing Bodies” was discussed the

most by national and regional organisations, with 123

occurrences, whereas “Remote Operator Roles and

633

Responsibilities” was discussed the least by national

and regional organisations, with 21 occurrences.

Figure 6. Main sub-themes emerged from the theme

Maritime Legislation for UARSVO (Source: Author).

4.2.1 Governing Bodies

As shown in Figure 7, the following governing

bodies were identified at macro, meso, and micro levels

for UARSVO. The meso level received the most

references concerning supporting organisations,

indicating that port authorities should require the

infrastructure and technology necessary to assume the

risk of handling a MASS vessel for port operations.

Participants also noted that Vessel Traffic Service (VTS)

plays a crucial role in MASS surveillance within their

designated controlled areas, relying on close

collaboration with the ROC to ensure the safe

coexistence of MASS with crewed vessels, alongside

providing assistance in managing cyber incidents.

Challenges regarding national organisations granting

MASS clearance to enter territorial waters when

operating in autonomous mode or under remote

control from another country were identified.

Participants emphasised the need for shipping

companies to designate a ROC in the country where the

vessel will primarily operate. Additionally, it was

asserted that this could facilitate access for the National

Accident Investigation Organisation (NAIO) to

Voyage Data Recording (VDR) data at the ROC of that

country. To support UARSVO accident investigations,

adherence to ANS's standards for explainability in

decision-making was deemed essential upon receipt of

VDR data. For this scenario, participants highlighted

the importance of close collaboration with technology

providers and during operations with the ROC for

troubleshooting purposes.

The second most referenced level was the macro.

Individuals expressed that the IMO is leading the

charge on international regulations with the MASS

Code, emphasising the need for close collaboration

with flag states, classification societies, and other

entities. There was a call for a new section on MASS to

be included in the International Convention on

Standards of Training, Certification and

Watchkeeping for Seafarers (STCW), similar to what

was added for STCW for Fishing Vessel Personnel

(STCW-F) [89]. Participants also raised concerns

regarding the contract of carriage and insurance

challenges that may arise if a software error from an

ANS is detected after the MASS vessel is delivered

from the shipyard to the ship owner. In this scenario,

the vessel enters the owner’s orbit, obligating them to

exercise due diligence in ensuring the vessel's

seaworthiness before and at the beginning of the

voyage. From a micro-level perspective, pilots were

regarded as essential for both arrival and departure

during UARSVO, providing critical local port

knowledge and preventing significant accidents (e.g.

the pilot saved lives by contacting from the container

ship Dali to halt traffic on the Francis Scott Key Bridge

(Baltimore, USA) prior to a potential collision with the

bridge) [90]. Participants also noted communication

challenges due to the vessel's remote command by the

master, suggesting that command of the vessel should

be transferred to the pilot during pilotage waters.

Figure 7. Operational Levels of Analysis for Uncrewed

Autonomous and Remote Surface Vessel Operations

(UARSVO) (Source: Author).

4.2.2 Remote Operator Roles and Responsibilities

The roles and responsibilities of RO most frequently

mentioned by participants were emergency handling,

passage planning, and cyber security for systems

assurance as primary roles, with special emphasis on

maintaining internal communication within ROC

personnel and the company and external

communication with other stakeholders. Other roles

and responsibilities listed by the frequency of

references included collision avoidance, preventive

maintenance, cargo operations, ship stability, ship

handling, and alternative fuel awareness. Participants

noted that these roles and responsibilities could vary

depending on the level of autonomy. It was also

mentioned that these could depend on the capabilities

of each shipping company's ROC, such as following the

sun operations, manning of ROC personnel, vessel

operation types, vessel types, the number of vessels

operating from an ROC, and whether a specialist RO

was created depending on the application.

The roles and responsibilities of the MASS Master

were highlighted as likely to shift to a more specialised

role with multi-vessel responsibilities for vessels

operating in the open sea. Participants expressed the

need to define the number of vessels a Master could be

in command of, in which circumstances, and also the

required backup Master(s) needed for redundancies.

For operations in which the ANS is with the conduct of

the MASS vessel, participants expressed that the

Master should be responsible for the vessel but should

not have full command responsibility as it may not be

fully aware of the MASS vessel situation. Additionally,

as the operation will be heavily reliant on technology,

the Master was also seen as responsible for the

performance of the systems and cyber security.

634

4.2.3 Remote Operator Education and Training

Standards

Participants' most referenced topics within the RO

education and training standard were the need for

training standardisation around technology operated

from ROC and cyber security of that technology as new

education, training and awareness on digital systems.

Additionally, special training was suggested for

scenarios where latency and connectivity can be

experienced from ROC, SA acquisition from

navigational equipment operated from ROC, and

system diagnosis for equipment fault. Along with this,

participants stated the shift needed for RO candidates

to attain training on electronic navigation equipment

from the start and learn traditional negational methods

as a backup. Individuals perceived that ROs would

need similar training to VTS Officers as they will be

required to have similar rational critical thinking,

shared decision-making, cognitive ability to focus and

duty duration. The importance of ROs' mental well-

being was also highlighted concerning the risk of social

life side effects when performing remote operations, as

ROs will encounter shoreside social challenges (no

longer working in isolation at sea). From the

participants' perspective, training will be necessary to

fulfil the roles and responsibilities listed in 4.2.2.

Participants stated that METI’s responsibility would be

to equip ROs with transferable skills, enabling this

workforce to become more employable, resourceful,

and adaptable, as current mariner training was

perceived to lack transferable skills ashore.

Concerning the training methods required for ROs,

the most frequent response was the advantage of

onboard mariner training for remote operations, as it

imparts essential seamanship skills.

Participants deemed it safest to implement short-

and medium-term ROs with sea time experience

(mariners holding a CoC transitioning to become an

RO). After a training maturity was attained and based

on the experience gained from the training provided

for those ROs, opportunities for RO training without

sea time experience were deemed as an opportunity to

be established in the long term. To achieve this, a

strong emphasis was placed on establishing simulator-

based training that could offer meaningful sea time in

a safe environment for the roles and responsibilities

required from ROs at the ROC. This includes

reinforcement and refresher training for ROs, as well

as a new type of training focused on how to interact

with an ANS. Furthermore, individuals from

company-based training viewed it as essential for

organisations with an ROC to incorporate simulator-

based training within their own ROC frameworks,

providing a standardised familiarisation for ROs with

the necessary time to develop remote situational

awareness from the ROC. Alternatively, the use of

Virtual Reality (VR) was mentioned to achieve a

general experience and feel of real-world scenarios, but

for limited periods of time, as some participants

experienced nausea.

4.2.4 Technology and Operational Standards for HAT

The topic that received the most occurrence within

this sub-theme was the standardisation and

certification of safety assurance of autonomy and ROC.

Individuals seem essential to attain a fail-safe

mechanism for MASS when operating outside the

operational design domain, such as dropping the

anchor for emergency anchorage, stopping the vessel

to hold position, assuming remote control by RO or

performing a hot swap with another vessel.

Additionally, redundancies for a backup ROC,

communication channels and navigation equipment

made by a manufacturer were stated. For this,

participants perceived that a standardisation of ANS

alarms and indications for RO would be needed to be

created in combination with an agnostic ROC user

interface for visuals (e.g. when multiple ROCs are

operating the same vessel) and a bird's eye view set up

from ROC (rather than traditional bridge ship

view). This extended to the MASS safety and

operational procedures with the need for a

standardised risk assessment similar to the System

Theoretic Process Analysis (STPA) approach or Risk

Based Assessment Tool (RABAT) by the European

Maritime Safety Agency (EMSA) [91, 92].

The second-most frequently discussed topic was the

standardisation of industry best practices, such as

seamless handovers for the command of a MASS vessel

between ROCs on international operations,

standardisation of communications, manning of ROCs

for the different phases of the voyage, and minimum

frequency of training for RO. Alternatively, new

standards for expansion and clarification of the IMO

degrees of autonomy were requested to define clear

roles and responsibilities between humans and

autonomy to acquire an international unified

taxonomy and terminology for maritime autonomy.

4.3 Maritime Cyber Security Resilience in UARSVO

Figure 8 demonstrates that of the four sub-themes

within the theme “Maritime Cyber Security Resilience

in UARSVO”, “HAT Strategies for Maritime Cyber

Security Resilience” was discussed the most, with 444

occurrences; on the other hand, “HAT Enhanced Cyber

Security Resilience Standards” was discussed the least,

with 67 occurrences. Furthermore, amongst key groups

of MASS stakeholders, “HAT Strategies for Maritime

Cyber Security Resilience” was discussed the most by

industrialists and resource providers, with 86

occurrences, whereas “Consequences of a Cyber

Incident in UARSVO” was discussed the least by

shipping companies and future workforce, with 10

occurrences.

Figure 8. Main sub-themes emerged from the theme

Maritime Cyber Security Resilience in UARSVO (Source:

Author).

635

4.3.1 Cyber-incidents in UARSVO

Within the sub-theme of cyber incidents in

UARSVO, participants' most frequently referenced

spoofing of the MASS vessel's Global Navigation

Satellite System (GNSS) and jamming of GNSS. For the

secondary category, the most mentioned is an incident

caused accidentally due to personnel incompetence,

ransomware and Automatic Identification System

(AIS) spoofing. The third category encompassed

Denial of Service (DDoS) attacks, man-in-the-middle

attacks, adversarial AI, social engineering, and insider

threats.

Participants noted that the primary threat actors

they believed responsible for these cyber incidents

were predominantly nation-states, owing to their

access to resources and expertise in the maritime field,

which enables sophisticated cyber operations. The next

most mentioned actors were terrorists and pirates, who

tend to engage in less sophisticated cyber activities.

Others cited include activists, criminals, and those

involved in business espionage. Participants indicated

that the principal motivations behind these cyber

incidents were likely financial gain and reputational

damage.

4.3.2 Consequences of a Cyber Incident in UARSVO

All except for one participant indicated that the

consequences of a cyber incident in UARSVO would be

either the same or higher. Participants' major concern

was that a cyber incident could affect critical

infrastructure due to gaining control of the MASS

vessel steering and propulsion system; such examples

given were a MASS vessel colliding with another vessel

or infrastructure, grounding, blocking a critical choke

point, causing environmental damage, and supply

chain disruptions. The secondary consequences were

company data loss and loss of property, life, and

reputation. Other less frequently mentioned were

cyber incidents affecting port Operational Technology

(OT) and national security. Amongst the higher

consequence group, participants perceived that a cyber

incident compromising ROC could affect an entire

company’s fleet and the worldwide economy due to

the connectivity of operations. Conversely, only one

participant opposed, arguing that the consequences are

lower as there could be no loss of life or environmental

damage scenarios, considering there may not be

anyone onboard and that alternative fuel or electric

batteries may already be in use.

4.3.3 HAT-Enhanced Cyber Security Resilience

Standards

Among the standards and frameworks for maritime

cyber resilience, the primary ones mentioned by

participants for continuity were the Unified

Requirements (UR) established by the International

Association of Classification Societies (IACS) alongside

the E26 (Guidelines for Cyber Resilience of Ships) and

E27 (Guidelines for Cyber Resilience of On-board

Systems and Equipment), which began to be applied to

new ships from 1 January 2024. Additionally, the

International Organization for Standardization

(ISO)/International Electrotechnical Commission (IEC)

27001 standard for Information Security Management

Systems (ISMS) was noted, along with the NIST

framework adopted by the IMO. Furthermore,

participants indicated that greater emphasis should be

placed on the standardisation of VDRs in UARSVO, as

this would assist in investigating causal factors for

cyber incidents.

Participants stated that training should be

integrated from the beginning and embedded within

the secure-by-design implementation for MASS as

HAT-enhanced cyber security resilience standards.

Participants perceive a need for the standardisation of

training in maritime cyber security resilience at the

international level, enabling nations to ratify and adopt

international standards to their own national

guidelines, thereby ensuring the safety and security of

MASS vessels. To facilitate this, individuals believe

that a gradual introduction is essential for

progressively assimilating the new roles and

responsibilities that come with maturity.

4.3.4 HAT Strategies for Cyber Security Resilience

The primary strategy highlighted by participants

and analysed by the authors for HAT cyber security

resilience in UARSVO was centred on system

resilience, particularly through the provision of

redundancies with other instruments during non-

nominal operations when systems do not function as

anticipated (such as using an Inertial Navigation

System (INS) and the next generation of AIS which is

VHF Data Exchange System (VDES)) [93, 94]. Sensor

fusion was regarded as the most effective approach for

achieving this, with the potential to employ an

intrusion detection system integrated with AI to

identify any cyber incidents, thereby minimising

human error in such situations.

Participants also noted that a fail-safe mechanism

should be established for MASS during a cyber

incident, ensuring the redundancies mentioned earlier

in 4.2.4 Technology and Operational Standards for

HAT. Furthermore, individuals emphasised the

necessity of new alert systems for cyber incidents

regarding ROC, the ROC user interface, and the

transparency of communication in decision- making

from ANS to RO concerning the information gathered

for SA and decision-making, which will be essential for

verifying the integrity of systems and ensuring data

quality. Additionally, individuals stated that a zero-

trust architecture should be implemented for

UARSVO, even though there are currently no ISO

standards for this.

The second most significant strategy referenced was

the Cyber Security Management System (CSMS) at the

company level. The area most frequently identified as

needing attention within the CSMS was the response

procedure from the ROC during a cyber incident. This

includes the procedure for assuming manual control of

the MASS vessel, if necessary, during a work case

scenario, as well as the decision-making process to

follow in the event of a cyber incident. Participants also

emphasised that a culture of cyber resilience should be

embedded as part of organisational strategy, alongside

risk assessments that incorporate multi-level

safeguarding and a dedicated response system for

cyber incidents to provide optimal assistance when

required. Consequently, they emphasised the

importance of having a checklist for work case cyber

incident scenarios and conducting regular audits and

assessments of these.

636

The third highest strategy mentioned was HAT

training in maritime cyber security resilience for

UARSVO, which is closely linked to the previous two.

Individuals highlighted the importance of training in

distinguishing between cyber-attacks and normal

system failures (as this is currently a challenge faced

within the maritime industry) and training for

maximum cyber incident crisis scenarios. It was

suggested that the development of these should be

undertaken by flag states and classification societies

but designed by the IMO. Additionally, the embedding

of cyber-attacks as an emergency was proposed.

The fourth-mentioned strategy was a socio-

technical systems approach from a multi-level

perspective. Participants asserted that this could help

strengthen the relationship between stakeholders

during cyber incidents, such as exploring the best

approaches in various scenarios to identify which team

and individual are best suited to assume responsibility

in a cyber incident.

Individuals noted that this could enhance

information sharing and lessons learned across the

maritime industry, such as developing a cohesive

paperless report that is valuable across all technology

suppliers and stakeholders, like the maritime single

window established by the IMO from January 2024 for

ships when entering a port but applied to cyber

incidents [95].

4.4 Maritime Cyber Security Resilience Training for

UARSVO

Figure 9 demonstrates that of the four sub-themes

within the theme “Maritime Cyber Security Training

for UARSVO”, “HAT Training and Procedures for

Cyber Security in UARSVO” was discussed the most,

with 140 occurrences; on the other hand, “RO HAT

Cyber Resilience Training for UARSVO” was

discussed the least, with 38 occurrences. Furthermore,

amongst key groups of MASS stakeholders,

“Organisational HAT Cyber Resilience Training for

UARSVO” and “HAT Training and Procedures for

Cyber Security in UARSVO” were discussed the most

by research and academia, with 32 occurrences,

whereas “RO HAT Cyber Resilience Training for

UARSVO” was discussed the least by international

organisations, with 1 occurrence.

Figure 9. Main sub-themes emerged from the theme

Maritime Cyber Security Resilience Training for UARSVO

(Source: Author).

4.4.1 RO HAT Cyber Resilience Training for UARSVO

The most frequently reported topic within the RO

HAT cyber resilience training was RO training on

recognising and managing a cyber incident.

Participants noted that these RO operators would play

a crucial role, acting as the first human line of defence

during a cyber incident, which may or may not involve

additional collaboration with personnel from the ROC

or external stakeholders. To address this, cyber

awareness training and education were seen as

needing different approaches for individuals with sea

time experience (traditional mariners) transitioning to

RO roles compared to those ROs lacking real sea time

experience. Furthermore, participants indicated that

ROs would need to learn cyber terminology to

communicate with cyber analysts effectively and

should grasp how systems operate to gain a

foundational understanding that could assist the RO

during a cyber incident.

4.4.2 Organisational HAT Cyber Resilience Training for

UARSVO

Training at the operational level was the most

frequent occurrence at the organisational level. There

was a clear need among participants for a maritime

cyber security expert or a technical operator who could

work alongside or embedded within the navigation

team in the ROC as individuals perceived challenges in

combining the RO role with the cyber analyst.

Moreover, participants emphasised that cyber analysts

and the chief engineer of the MASS vessel will need to

work closely with the RO and the ROC management

during a cyber incident. Within the ROC for a cyber

incident, two approaches were seen as viable for

shipping companies with an ROC to adopt: either

setting up a cyber crisis management team that

constantly operates or key people among different

departments forming the cyber crisis management

team.

At the management level, participants believed that

ROC management should require both operational and

technical skills to effectively handle the responsibilities

during a cyber incident, as previously mentioned. The

ROC manager was viewed as closely collaborating

with the organisation's fleet manager, Company

Security Officer (CSO), and Designated Person Ashore

(DPA). In some scenarios, as indicated by participants,

the Security Operations Centre (SOC) (which could be

within the ROC or third party) must feed that

information to the ship owner and MASS Master. If it

fails to, the SOC might determine that the MASS vessel

is not seaworthy without the Master knowing it.

Therefore, the right information and training would

need to be provided for the Master so that they can take

the appropriate action when detecting a cyber-attack,

as they are ultimately responsible for the command of

the MASS vessel. Participants stated that in the UK,

shipping companies are considered Operator of

Essential Services (OES) under Network and

Information Systems (NIS) regulation and that for

incidents meeting the mandatory reporting thresholds,

such as hampering the ability to deliver essential

services (e.g. closing ports across the UK) should report

to the Department for Transport (DfT) Cyber

Compliance Team (CCT) no later than 72 hours (as per

Part 3 of the Data Protection Act 2018) of the notifiable

637

incident has happened [96-98]. Consequently,

individuals stated that it ultimately depends on the

company’s board-level culture whether the ship owner

takes responsibility for addressing the cyber incident,

which could potentially lead to micromanagement or

whether this responsibility is delegated to technical

management. If full trust is placed in the management

team, the ship owner will be distanced from the cyber

incident.

For this, training and standardised procedures to

follow by the ROC team during a cyber incident were

seen as essential, since the ability to quantify non-

nominal operations uncertainty and how to act upon

those uncertainties with tools such as checklists and

up-to-date reporting records.

4.4.3 Multi-organisation HAT Cyber Resilience Training

for UARSVO

Within multi-organisation HAT cyber resilience

training, the meso level received the most occurrences,

especially regarding technology suppliers. These

organisations have encountered different challenges

that could cause a cyber incident, incorrect system

integration, improper use of equipment or another

cause and perceived the need to receive a reported

tracking back with how systems are operated and

maintained. As vessels operate with many equipment

manufacturers, they perceived that a single point of

contact from all the Original Equipment Manufacturers

(OEM) vendors would be needed for UARSVO to

facilitate efficiency during cyber incident response

collaboration with ROC. Other stakeholders were also

regarded as directly involved during a cyber incident,

such as the VTS, coastal authorities and port authorities

intervening to help ensure the safety of the vessel in the

vicinity of the designated area and providing backup

assistance to the ROC operating the MASS vessel.

Alternatively, NAIOs are seen as advantageous for

collecting data on cyber incidents to provide the

industry with valuable lessons from various scenarios,

as in some cases, participants stated that NAIOs may

be the first to identify that the cause of an accident was

a cyber incident.

At the macro level, the insurance perspective for

training during a cyber incident on UARSVO was seen

as essential to proactively look at preventive measures,

such as the provided identification of roles and

responsibilities of the Company Financial Officer

(CFO), negotiators, and lawyers during a cyber

incident as well as coverage under the premium

insurance for a cyber incident. Alongside this

classification societies were seen as essential with the

smart notations implemented for ships that use smart

technology, which helps detect early any potential

failure. At the micro level, participants stated that each

micro-specialist during a cyber incident could assist in

providing crucial information such as marine pilots,

law enforcement entities and terminal operators.

Consequently, individuals suggested the

implementation of joint multi-disciplinary training

exercises for cyber-incidents to help define roles and

responsibilities internally and externally of the

organisations and attain the needed joint experiential

learning within ROC and UARSVO stakeholders.

4.4.4 HAT Training and Procedures for Cyber Security

in UARSVO

For this sub-theme, the implementation of cyber

security resilience training in UARSVO has been the

most frequently mentioned. Participants indicated that

current training and education on the latest cyber

threat landscape should be provided to the maritime

workforce. Individuals believe that a foundational

level of training and awareness on maritime cyber

security has been delivered at all competency levels in

recent years since the release of Resolution

MSC.428(98) on Maritime Cyber Risk Management in

Safety Management Systems [41]. However,

specialised training that links the decision-making

process across macro, meso, and micro hierarchy levels

is necessary for high-risk cyber incident scenarios that

escalate through various operational levels of analysis.

Participants requested additional training focused on

critically assessing information provided by systems,

incorporating real-world scenarios, to establish an

innovative training response framework that can be

tested with maritime personnel to observe the

differences in responses between those with and

without maritime cyber security training. Participants

also expressed that regular training would be essential

to acquire the experience and knowledge required to

respond to unforeseen cyber incident scenarios that

may arise within UARSVO. The frequency of training

was emphasised to be tailored to each organisation's

needs to ensure cost efficiency. Furthermore,

individuals noted that research and academia could be

the initial implementers of this training, after which a

third-party training company could expand and

develop customised solutions suitable for

organisations.

The most frequently referenced training methods

for UARSVO were the use of simulators, which

provide a safe environment for training, and tabletop

exercises, as these are straightforward to conduct [99].

Either method was considered valuable depending on

the ROC organisation and the type of vessel operation.

In addition, participants with a procurement cycle

perspective suggested an objective-based assessment

approach for determining training requirements when

introducing this training. Moreover, participants

suggested using an objective-based approach to

wargaming for this training, rather than one centred on

procedural risks, to ensure it is goal-oriented instead of

prescriptive about the procedures, as it offers flexibility

for targeted outcomes.

5 DISCUSSION AND FINDINGS

The ongoing development of UARSVO requires

multidisciplinary collaboration to address all

perspectives essential for ensuring a successful

transition when introducing new technology, as noted

in Section 4.2.1. This has been evident through the

engagement of key MASS stakeholders in this study,

particularly among governing bodies. As discovered

from Section 4.1.1 and 4.1.2, major advancements in

this development have been seen across developed

countries and regions where technology and

infrastructure have been well established to facilitate

these operations. Although a mandatory MASS Code

is aimed to be implemented in 2032 by the IMO, MASS

638

vessel acceptance will be dependent on the above, as

well as regulations established nationally and

regionally, having the ability only to perform

autonomous vessel navigation on dedicated sea areas

[1, 100, 101]. This also correlates with the results

provided in Figure 4, as a higher occurrence in the sub-

theme challenges for UARSVO was received across

international organisations, and in the sub-theme

opportunities for UARVO, it was received across

national organisations.

As stated by [102] and [103] due to the

interrelatedness in definitions between the terms

automation and autonomy, crucial awareness should

be provided for these terminologies, as the autonomy

of technology occurs when technology assumes

responsibility and makes decisions on behalf of

humans. This distinction is clearly illustrated in Section

4.1.3, obtained from participant contributions, and

highlights the need for the IMO to define the

terminology and taxology for this. Furthermore, the

HAT within the HAC requires a relationship similar to

that highlighted by participants and literature, viewing

technology at the same level as human-to-human

interactions rather than as mere tools [104, 105]. This

correlates with the [6] and [106] statement that the role

of the human element is perceived to evolve and

transition towards a more supervisory one.

Participants noted with the use of the interview

guide, as shown in Section 4.2.2, that RO roles and

responsibilities must be specialised in emergency

handling, passage planning, and cyber security for

systems assurance. Additionally, in Section 4.2.3 on RO

Education and Training Standards, training with

technology operated from ROC and cyber security

ranked among the two highest-referenced areas due to

the necessity of systems assurance. In each case, a

specialised RO role in these areas was seen as feasible,

depending on the needs of the ROC organisation, as

participants found it difficult to combine the

RO role with that of the cyber analyst. This

incorporation of cyber security for MASS has also been

regarded as essential to include as a new topic within

the maritime curricula for METIs, according to [107]

which involved 70 participants from various

stakeholder groups in the maritime sector.

Furthermore, the review provided by [108] on the IMO

Human Element, Training and Watchkeeping (HTW)

10th session in February 2024, stated the need for

amendments to incorporate hands-on experience with

simulations and practical exercises to enhance cyber

security awareness and training among mariners

within international maritime education and training

for crewed vessels. However, this remains

unaddressed as no amendments were made at the 11th

HTW held in February 2025 [109].

Participants perceived the development of training

on HAT for maritime cyber security resilience in

UARSVO as vital as ROs will serve as a crucial line of

defence in cyber incident management within ROC,

collaborating with other internal and external

personnel within and outside of the organisation. The

proposed approach for the short-term development

and implementation of this training was identified for

mariners with sea time experience transitioning into an

RO role, which will require future adaptation in the

long-term if ROs without sea time experience become

a reality in future UARSVO. Although HAT training

for maritime cyber security resilience was ranked as

the third strategy (see Table 5) by participants for HAT

in maritime cyber security resilience for UARSVO,

workforce training remains a vital organisational asset,

as it interrelated to the first two strategies, system

resilience and CSM in the event of a cyber incident.

Alternatively, the fourth most frequently

mentioned strategy was the socio-technical system

approach from a multi-level perspective during cyber

incident management, aimed at improving safety in

UARSVO [110, 111]. This also aligns, as highlighted in

Section 4.4.4, with the necessity of providing holistic,

specialised training that encompasses the connections

within the decision-making process across macro,

meso, and micro hierarchy levels during cyber incident

management, which escalates through various

operational levels analysis. [55] proposed future work

to develop a maritime cyber risk management training

framework specifically for crewed vessels, utilising the

HCD process at a holistic macro level. However, within

the maritime industry, the only standardised cyber

incident response reporting procedure varies from

country to country, and there is yet to be a standardised

procedure at the international level [96-98].

Table 5. Strategies that could best enhance HAT in maritime

cyber security resilience for UARSVO as per frequency

received by participants (source: author).

Strategies

Sub-Strategies Developed Developed

System

resilience

Redundancies with other instruments during non-

nominal operations

Sensor fusion

Fail-safe mechanism

Alert systems for cyber incidents regarding ROC

Transparency of communication in decision-

making from ANS to RO

Zero trust architecture

Cyber Security

Management

System (CSMS)

Response procedure from the ROC during a cyber

incident

Culture of Cyber Resilience

Risk assessments that incorporate multi- level

safeguarding

Dedicated response system for cyber incidents

Checklist for work case cyber incident scenarios

HAT training in

maritime cyber

security

resilience for

UARSVO

Training in distinguishing between cyber- attacks

and normal system failures

Training for maximum cyber incident crisis

scenarios

Embedding of cyber-attacks as an emergency was

proposed

Socio- technical

systems

approach from

a multi- level

perspective

Best approaches in various scenarios

Identify which team and individual are best suited

to assume responsibility in a cyber incident

Enhance information sharing and lessons learned

across the maritime industry

6 LIMITATION AND FUTURE WORK

The methodological implications of conducting semi-

structured interviews for this research provided a

holistic foundational knowledge and strategies for

HAT in maritime cyber security resilience for

UARSVO, building upon previous work that focused

solely on mariners [13]. However, the methodology

adopted rendered various challenges to obtaining the

data for the conduct of this study. One of the main

challenges was that the data acquisition from semi-

structured interviews was highly time-consuming due

to the time required to arrange interviews (which

639

included dropouts or the need to reschedule some

interviews) and the process of verifying transcriptions

and coding. The second main challenge was data

quality assurance, as the questions given were open-

ended, meaning that although this gave room to

explore and develop new themes, there was room for

proving generic answers. The answers from

participants also depended on their willingness and

knowledge to provide their perception and the order in

which longer answers were given, especially for the

first question set. Additionally, when interpreting the

data, this was subjective to the interview guide. During

this research, the interview process was stopped when

data saturation and consistency from the different

stakeholder groups were achieved. Other limitations

extended to attaining confidential or classified

information for each organisation. Although a

participant with prior experience in VTS, a participant

currently working in coastal administration, and

experienced ROs were interviewed (including

fieldwork with ROCs and discussions with currently

operating ROs), a broader diversity of opinions would

be achieved if currently active Vessel Traffic Service

(VTS) officers, coastal authorities, and ROs could also

have participated.

Future efforts could consider utilising the

comprehensive, in-depth foundational qualitative data

gathered from semi-structured interviews conducted

with key MASS stakeholders. In combination with

results obtained from [13], participant engagement

could be extended (as stated above) to develop and

validate a training framework for HAT in maritime

cyber security resilience for UARSVO. The framework

could then be tested and evaluated with key MASS

stakeholders in a workshop completing the iteration

steps from the hybrid DRM and HCD methodology

adopted.

7 CONCLUSION

This research, based on semi-structured interviews

with 76 participants, is the first of its kind due to the

diversity of industry experts who participated from

different key MASS industry stakeholder groups. The

findings were unique as these address a

comprehensive overview and foundational knowledge

of UARSVO, maritime legislation surrounding

UARSVO, maritime cyber security resilience related to

UARSVO, and training in maritime cyber security

resilience for UARSVO. Furthermore, the study

developed key strategies that could effectively enhance

HAT in maritime cyber security resilience for UARSVO

based on thorough expertise that participated in this

study from multi- disciplinary backgrounds within

maritime autonomy.

The first ranked proposed strategy focused on

systems resilience; the second emphasised Cyber

Security Management Systems (CSMS) at the company

level for ROCs; the third involved HAT training in

maritime cyber security resilience for UARSVO with

special focus on training to differentiate between

cyber-attacks and normal system failures, along with

training for maximum cyber incident crisis scenarios;

the fourth adopted socio-technical systems approach

from a multi-level perspective. Based on these findings,

while HAT training in maritime cyber security

resilience for UARSVO was the third most frequently

mentioned strategy, it is the one that underpins the first

two and the fourth during cyber incident management.

Furthermore, it was identified that within HAT cyber

resilience training for ROs will prioritise recognising

and managing a cyber incident as a primary focus for

training.

Results highlighted that future education and

training standards would require ROs to integrate

cyber security into METIs, as this would be one of the

primary roles and responsibilities of these ROs, even

though it is perceived to differ from that of a cyber

security expert. Additionally, the results offer a

comprehensive foundational knowledge that could

inform the development of future training for ROs in

HAT related to maritime cyber security resilience for

UARSVO. This could then serve as an interim training

framework under the IMO’s sub-committee on HTW,

enabling academia to develop training programmes on

cyber resilience for UARSVO until mandatory

instruments are established.

ACKNOWLEDGEMENTS

We would like to express our gratitude to all the interview

participants for their crucial contributions to this study.

Without them, the study could not have been conducted. We

extend our gratitude to the sponsors of this research, Reardon

Smith Nautical Trust and Cyber-SHIP lab, for enabling this

research.

REFERENCES

[1] IMO. "MSC 109/WP.8 Development of a Goal-Based

Instrument for Maritime Autonomous Surface Ships

(MASS) - Report of the MSC-LEG-FAL Joint Working

Group on Maritime Autonomous Surface Ships (MASS)."

IMO. https://docs.imo.org/ (accessed 21/01/2025, 2025).

[2] T. Johansen and I. B. Utne, "Human-autonomy

collaboration in supervisory risk control of autonomous

ships," Journal of Marine Engineering & Technology, vol.

23, no. 2, pp. 135-153, 2024/03/03 2024, doi:

10.1080/20464177.2024.2319369.

[3] K. Tam, "MASS Cybersecurity," in Maritime Autonomous

Surface Ships (MASS) - Regulation, Technology, and

Policy: Three Dimensions of Effective Implementation,

C.-J. Chae and R. Baumler Eds. Cham: Springer Nature

Switzerland, 2024, pp. 165-180.

[4] DNV. "DNV-RP-0323 - Certification scheme for remote

control centre operators." DNV. https://rules.dnv.com/

(accessed 21/01, 2025).

[5] DNV. "DNV-ST-0324 - Competence of remote control

centre operators." DNV. https://rules.dnv.com/ (accessed

21/01, 2025).

[6] J. Chan, D. Golightly, R. Norman, and K. Pazouki,

"Perception of Autonomy and the Role of Experience

within the Maritime Industry," Journal of Marine Science

and Engineering, vol. 11, no. 2, doi:

10.3390/jmse11020258.

[7] J. Y. C. Chen, A. R. Selkowitz, K. Stowers, S. G. Lakhmani,

and M. J. Barnes, "Human-Autonomy Teaming and

Agent Transparency," presented at the Proceedings of the

Companion of the 2017 ACM/IEEE International

Conference on Human-Robot Interaction, Vienna,

Austria, 2017. [Online]. Available:

https://doi.org/10.1145/3029798.3038339.

[8] K. T. Wynne and J. B. Lyons, "An integrative model of

autonomous agent teammate-likeness," Theoretical

640

Issues in Ergonomics Science, vol. 19, no. 3, pp. 353-374,

2018/05/04 2018, doi: 10.1080/1463922X.2016.1260181.

[9] T. O’Neill, N. McNeese, A. Barron, and B. Schelble,

"Human–Autonomy Teaming: A Review and Analysis of

the Empirical Literature," Human Factors, vol. 64, no. 5,

pp. 904-938, 2022/08/01 2020, doi:

10.1177/0018720820960865.

[10] T.-E. Kim, L. P. Perera, M.-P. Sollid, B.-M. Batalden, and

A. K. Sydnes, "Safety challenges related to autonomous

ships in mixed navigational environments," WMU

Journal of Maritime Affairs, vol. 21, no. 2, pp. 141-159,

2022, doi: 10.1007/s13437-022-00277-z.

[11] M. Walter, A. Barrett, D. Walker, and K. Tam,

"Adversarial AI Testcases for Maritime Autonomous

Systems," AI, Computer Science and Robotics

Technology, vol. 2, 04/25 2023, doi: 10.5772/acrt.15.

[12] DNV. "Tackling a growing cybersecurity threat in an

increasingly connected industry." DNV.

https://www.dnv.com/expert-story/maritime-

impact/tackling-a-growing-cybersecurity-threat-in-an-

increasingly-connected-

industry/#:~:text=Scale%20and%20frequency%20of%20c

yber,over%20the%20previous%20five%20y ears.

(accessed 21/01, 2025).

[13] J. D. Palbar Misas, R. Hopcraft, K. Tam, and K. Jones,

"Future of maritime autonomy: cybersecurity, trust and

mariner's situational awareness," Journal of Marine

Engineering & Technology, pp. 1-12, 2024, doi:

10.1080/20464177.2024.2330176.

[14] V. Bolbot, K. Kulkarni, P. Brunou, O. V. Banda, and M.

Musharraf, "Developments and research directions in

maritime cybersecurity: A systematic literature review

and bibliometric analysis," International Journal of

Critical Infrastructure Protection, vol. 39, p. 100571,

2022/12/01/ 2022, doi:

https://doi.org/10.1016/j.ijcip.2022.100571.

[15] N. Tabish and T. Chaur-Luh, "Maritime Autonomous

Surface Ships: A Review of Cybersecurity Challenges,

Countermeasures, and Future Perspectives," IEEE

Access, vol. 12, pp. 17114-17136, 2024, doi:

10.1109/ACCESS.2024.3357082.

[16] E. Veitch, A. Hynnekleiv, and M. Lützhöft, "The

Operator's Stake in Shore Control Center Design: A

Stakeholder Analysis for Autonomous Ships,"

International Conference on Human Factors 2020, 02/19

2020, doi: 10.3940/hf.20.5.

[17] A. Nuwan, M. Shukri, and A. Khatibi, "Managerial

Strategies for Organisational Resilience in SME Sector,"

International Journal of Research and Scientific

Innovation, vol. xi, pp. 384-404, 01/01 2024, doi:

10.51244/IJRSI.2024.1101030.

[18] E. Erstad, "Operational training for enhanced maritime

cyber resilience: Bridging safety and security through

maritime education and training," Doctoral theses,

Department of Ocean Operations and Civil Engineering,

NTNU, NTNU Open, 2024:109, 2024. [Online]. Available:

https://ntnuopen.ntnu.no/ntnu-

xmlui/handle/11250/3125566

[19] Maritime Coastguard Agency. "Maritime Autonomy

Regulation Lab (MARLab) Report." Gov.UK.

https://www.gov.uk/government/publications/maritime-

autonomy-regulation-lab-marlab- report (accessed 07/02,

2025).

[20] AUTOSHIP. "D3.1 - Autonomous ship design

standards." European Comission.

https://ec.europa.eu/research/participants/documents/do

wnloadPublic?documentIds=080166e5da

3b7aa6&appId=PPGMS (accessed 18/02, 2025).

[21] T. Relling, M. Lützhöft, R. Ostnes, and H. P. Hildre, "A

Human Perspective on Maritime Autonomy," in

Augmented Cognition: Users and Contexts, Cham, D. D.

Schmorrow and C. M. Fidopiastis, Eds., 2018// 2018:

Springer International Publishing, pp. 350-362.

[22] Cambridge Dictionary. "Definition of Automation."

https://dictionary.cambridge.org/dictionary/english/auto

mation (accessed 22/01, 2025).

[23] L. Bainbridge, "Ironies of automation," Automatica, vol.

19, no. 6, pp. 775-779, 1983/11/01/ 1983, doi:

https://doi.org/10.1016/0005-1098(83)90046-8.

[24] N. Sarter, D. Woods, and C. Billings, "Automation

surprises," vol. 2, 1997, pp. 1926-1943.

[25] National Transport Safety Board. "Collision between US

Navy Destroyer John S McCain and Tanker Alnic MC

Singapore Strait, 5 Miles Northeast of Horsburgh

Lighthouse." National Transport Safety Board.

https://www.ntsb.gov/investigations/accidentreports/rep

orts/mar1901.pdf (accessed 23/01, 2025).

[26] J. Rasmussen, "Risk management in a dynamic society: a

modelling problem," Safety Science, vol. 27, no. 2, pp. 183-

213, 1997/11/01/ 1997, doi: https://doi.org/10.1016/S0925-

7535(97)00052-0.

[27] J. Rasmussen, Information processing and human-

machine interaction : an approach to cognitive

engineering (North-Holland series in system science and

engineering ; 12). New York: North-Holland, 1986.

[28] D.-T. Pham, H. Ali, K. Fennedy, M.-H. Hsieh, S. Alam,

and V. Duong, "Human-AI Hybrid Paradigm for

Collaborative Air Traffic Management Systems."

[29] D. H. Hagos, H. El Alami, and D. B. Rawat, AI-Driven

Human-Autonomy Teaming in Tactical Operations:

Proposed Framework, Challenges, and Future Directions.

2024.

[30] Ø. Rødseth, L. A. Wennersberg, and H. Nordahl, "Levels

of autonomy for ships," Journal of Physics: Conference

Series, vol. 2311, p. 012018, 07/01 2022, doi: 10.1088/1742-

6596/2311/1/012018.

[31] European Maritime Safety Agency. "CMOROC Final

Report - Identification of Competences for MASS

Operators in Remote Operation Centres."

https://www.emsa.europa.eu/publications/reports/item/

5089-cmoroc-mass.html (accessed 23/01, 2025).

[32] Z. H. Munim, T. Notteboom, H. Haralambides, and H.

Schøyen, "Key determinants for the commercial

feasibility of maritime autonomous surface ships

(MASS)," Marine Policy, vol. 172, p. 106482, 2025/02/01/

2025, doi: https://doi.org/10.1016/j.marpol.2024.106482.

[33] Maritime UK. "Maritime Autonomous Ship Systems

(MASS) UK Industry Conduct Principles and Code of

Practice A Voluntary Code - Version 8 " The Society of

Maritime Industries (SMI).

https://www.maritimeindustries.org/application/files/94

17/3375/8340/MASS_COP_2024_V8_pages_V1.1.pdf

(accessed 14/02, 2025).

[34] (2018). Maritime Safety Committee - 100th Session,

Agenda Item 5: Regulatory Scoping Exercise for the use

of Maritime Autonomous Surface Ships (MASS).

[35] Bureau Veritas. "Guidelines for Autonomous Shipping."

Bureau Veritas.

https://erules.veristar.com/dy/data/bv/pdf/641-NI_2019-

10.pdf (accessed 22/01, 2025).

[36] DNV. "DNVGL-CG-0264 Autonomous and remotely

operated ships." DNV.

https://www.smashroadmap.com/files/dnvgl-cg-.pdf

(accessed 22/01, 2025).

[37] ONESEA. "AUTONOMOUS SHIPS TERMS OF

REFERENCE FOR RULE DEVELOPMENT." ONEDEA.

https://b307d3dfff.clvaw-

cdnwnd.com/ac8e62b9b6e49d3c818fcd9770855fb0/20000

0069-

3d56a3d56d/WHITEPAPER%202022_One%20Sea%20-

%20Terminology%20and%20levels%20of%20automation

.pdf?ph=b307d3dfff (accessed 22/01, 2025).

[38] Lloyds Register. "LR Code for Unmanned Marine

Systems." Lloyd’s

e_for_Unmanned_Marine_Systems__February_20172_.p

df (accessed 22/01, 2025).

641

[39] N. P. Ventikos. "Determining the requirements for fail-

safe design of autonomous ships: The MOSES

Autonomous Tugboat Swarm." https://moses-

h2020.eu/wp- content/uploads/2024/06/Autonomous-

tugboats-failsafe-approach_v1.0.pdf (accessed 28/01,

2025).

[40] DNV. "Maritime Cyber Priority 2024/25: Managing cyber

risk to enable innovation."

https://www.dnv.com/cyber/insights/publications/down

load/thank-you-download-2024-maritime- cyber-

priority/ (accessed 28/01, 2025).

[41] IMO. "Resolution MSC.428(98) – Maritime Cyber Risk

Management in Safety Management Systems " IMO.

https://wwwcdn.imo.org/localresources/en/OurWork/Se

curity/Documents/Resolution%20MSC.428( 98).pdf

(accessed 28/01, 2025).

[42] IMO. "MSC-FAL.1-Circ.3-Rev.2 - Guidelines On

Maritime Cyber Risk Management." IMO.

https://wwwcdn.imo.org/localresources/en/OurWork/Se

curity/Documents/MSC-FAL.1-Circ.3- Rev.2%20-

%20Guidelines%20On%20Maritime%20Cyber%20Risk%

20Management%20(Secretariat)%20(1).pdf (accessed

29/01, 2025).

[43] P. Peng, X. Xie, C. Claramunt, F. Lu, F. Gong, and R. Yan,

"Bibliometric analysis of maritime cybersecurity:

Research status, focus, and perspectives," Transportation

Research Part E: Logistics and Transportation Review,

vol. 195, p. 103971, 2025/03/01/ 2025, doi:

https://doi.org/10.1016/j.tre.2025.103971.

[44] NIST. "Cyber Resiliency Definition."

https://csrc.nist.gov/glossary/term/cyber_resiliency

(accessed 29/01, 2025).

[45] National Institutes of Health. "What are competencies?"

NIH. https://hr.nih.gov/about/faq/working-

nih/competencies/what-are-

competencies#:~:text=Knowledge%20is%20information

%20developed%20or,individual%20to%20a%

20certain%20situation. (accessed 29/01, 2025).

[46] MASSPeople. "MSC/ISWG/MASS 2/3/1 Consideration of

Proposals for Further Development of the Draft MASS

Code. Bridging Competency Gaps for MASS Operators in

Alignment with the STCW Framework." IMO.

https://www.masspeople.org/imo-submission/msc-iswg-

mass-2-3-1 (accessed 08- 03, 2025).

[47] R. Hopcraft, "Developing Maritime Digital

Competencies," IEEE Communications Standards

Magazine, vol. 5, no. 3, pp. 12-18, 2021, doi:

10.1109/MCOMSTD.101.2000073.

[48] C. Campos Toresano, M. L. Castells Sanabra, and C.

Borén Altés, "Conceptualization of the era of autonomous

shipping within maritime education and training

framework," 2024 2024: Universitat Politècnica de

Catalunya. Iniciativa Digital Politècnica, doi:

10.5821/mt.12808. [Online]. Available:

https://dx.doi.org/10.5821/mt.12808

[49] MAIB. "Collision between pure car carrier City of

Rotterdam and ro-ro freight ferry Primula Seaways."

GOV.UK. https://www.gov.uk/maib-reports/collision-

between-pure-car-carrier-city-of- rotterdam-and-ro-ro-

freight-ferry-primula-seaways (accessed 29/01, 2025).

[50] A. Landman, E. L. Groen, M. M. van Paassen, A. W.

Bronkhorst, and M. Mulder, "Dealing With Unexpected

Events on the Flight Deck: A Conceptual Model of Startle

and Surprise," Human Factors, vol. 59, no. 8, pp. 1161-

1172, 2017/12/01 2017, doi: 10.1177/0018720817723428.

[51] S. Symes, E. Blanco-Davis, T. Graham, J. Wang, and E.

Shaw, "The survivability of autonomous vessels from

cyber-attacks," Journal of Marine Engineering &

Technology, pp. 1-23, 2024, doi:

10.1080/20464177.2024.2428022.

[52] University of Plymouth. "Maritime-Cyber Ransomware

Scenario."

https://www.youtube.com/watch?v=WZvkWIp6sEQ

(accessed 30/01, 2025).

[53] University of Plymouth. "False AIS Data in Cyber-Attack

Scenario."

https://www.youtube.com/watch?v=fTjCr0X0nRQ

(accessed 30/01, 2025).

[54] NHL Stenden. "Maritime Cyber Incident Simulations."

https://www.nhlstenden.com/en/maritime-cyber-

incident-simulations (accessed 30/01, 2025).

[55] E. Erstad, R. Hopcraft, A. Vineetha Harish, and K. Tam,

"A human-centred design approach for the development

and conducting of maritime cyber resilience training,"

WMU Journal of Maritime Affairs, vol. 22, no. 2, pp. 241-

266, 2023/06/01 2023, doi: 10.1007/s13437-023-00304-7.

[56] D. Santos Castro, V. Plácido Da Conceição, and S.

Campaniço Cavaleiro, "PNT Resilience and the Impact of

Satellite Radio Positioning Disruptions on Piloting

Teams," Institute of Marine Engineering, Science and

Technology (IMarEST), 2022.

[57] Whiterbys, Cyber Security Workbook for On Board Ship

Use. London: Whiterbys, 2025, p. 238.

[58] E. Erstad, R. Hopcraft, J. D. Palbar, and K. Tam, "CERP:

A Maritime Cyber Risk Decision Making Tool,"

TransNav, the International Journal on Marine

Navigation and Safety of Sea Transportation, vol. 17, no.

2, pp. 269-279, 2023, doi: 10.12716/1001.17.02.02.

[59] A. Oruc, N. Chowdhury, and V. Gkioulos, "A modular

cyber security training programme for the maritime

domain," International Journal of Information Security,

vol. 23, no. 2, pp. 1477-1512, 2024/04/01 2024, doi:

10.1007/s10207-023-00799-4.

[60] A. Oruc, N. Chowdhury, V. Gkioulos, and S. Katsikas,

"Evaluation of Maritime Cyber Security (MarCy) Training

Programme," TransNav, the International Journal on

Marine Navigation and Safety of Sea Transportation, vol.

18, no. 4, pp. 743-763, 2024, doi: 10.12716/1001.18.04.01.

[61] A. Vineetha Harish, K. Tam, and K. Jones, "Generating

training events for building cyber- physical security

skills," The Computer Journal, p. bxae123, 2024, doi:

10.1093/comjnl/bxae123.

[62] S. M. Kim, "Inductive or deductive? Research by

maxillofacial surgeons," Journal of the Korean

Association of Oral and Maxillofacial Surgeons, vol. 47,

no. 3, pp. 151-152, 2021, doi: 10.5125/jkaoms.2021.47.3.151.

[63] A. Bryman and E. Bell, Business Research Methods.

Oxford: Oxford University Press, 2011.

[64] M. Saunders, P. Lewis, and A. Thornhill, Research

Methods for Business Students, 7th Edition ed. Harlow:

Pearson, 2016.

[65] J. W. Creswell and J. D. Creswell, Research Design:

Qualitative, Quantitative, and Mixed Methods

Approaches. Los Angeles: Sage, 2018.

[66] M. N. K. Saunders, P. Lewis, and A. Thornhill, Research

Methods for Business Students, 8th Edition ed. New

York: Pearson, 2019.

[67] W. M. Trochim and J. P. Donnelly, The Research

Methods Knowledge Base, 3rd ed. Cincinnati, OH:

Atomic Dog, 2006.

[68] L. T. M. Blessing and A. Chakrabarti, DRM, a Design

Research Methodology. Springer London, 2009, pp. XVII,

397.

[69] ISO. "9241-210: 2019 Ergonomics of human-system

interaction. Part 210: Human-Centred Design for

Interactive Systems. ." ISO.

https://www.iso.org/standard/77520.html (accessed.

[70] L. A. Palinkas, S. M. Horwitz, C. A. Green, J. P. Wisdom,

N. Duan, and K. Hoagwood, "Purposeful Sampling for

Qualitative Data Collection and Analysis in Mixed

Method Implementation Research," (in eng), Adm Policy

Ment Health, vol. 42, no. 5, pp. 533-44, Sep 2015, doi:

10.1007/s10488-013-0528-y.

[71] S. Sarantakos, Social Research, 3rd ed. London: Palgrave

Macmillan, 2005.

[72] C. Andrade, "The Inconvenient Truth About

Convenience and Purposive Samples," Indian Journal of

Psychological Medicine, vol. 43, no. 1, pp. 86-88,

2021/01/01 2020, doi: 10.1177/0253717620977000.

642

[73] S. Campbell et al., "Purposive sampling: complex or

simple? Research case examples," (in eng), J Res Nurs, vol.

25, no. 8, pp. 652-661, Dec 2020, doi:

10.1177/1744987120927206.

[74] G. R. Sadler, H. C. Lee, R. S. Lim, and J. Fullerton,

"Recruitment of hard-to-reach population subgroups via

adaptations of the snowball sampling strategy," (in eng),

Nurs Health Sci, vol. 12, no. 3, pp. 369-74, Sep 1 2010, doi:

10.1111/j.1442-2018.2010.00541.x.

[75] V. Braun and V. Clarke, Thematic Analysis A Practical

Guide, First ed. London: SAGE, 2022.

[76] M. M. Hennink, B. N. Kaiser, and V. C. Marconi, "Code

Saturation Versus Meaning Saturation:How Many

Interviews Are Enough?," Qualitative Health Research,

vol. 27, no. 4, pp. 591- 608, 2017, doi:

10.1177/1049732316665344.

[77] L. Nyanchoka, C. Tudur-Smith, R. Porcher, and D. Hren,

"Key stakeholders' perspectives and experiences with

defining, identifying and displaying gaps in health

research: a qualitative study protocol," (in eng), BMJ

Open, vol. 9, no. 8, p. e027926, Sep 3 2019, doi:

10.1136/bmjopen-2018- 027926.

[78] A. Galletta, Mastering the semi-structured interview and

beyond: From research design to analysis and publication

(Mastering the semi-structured interview and beyond:

From research design to analysis and publication.). New

York, NY, US: New York University Press, 2013, pp. xiii,

245-xiii, 245.

[79] N. King and C. Horrocks, Interviews in qualitative

research. London: Sage, 2010.

[80] B. Dicicco-Bloom and B. F. Crabtree, "The qualitative

research interview," (in eng), Med Educ, vol. 40, no. 4, pp.

314-21, Apr 2006, doi: 10.1111/j.1365-2929.2006.02418.x.

[81] M. DeJonckheere and L. M. Vaughn, "Semistructured

interviewing in primary care research: a balance of

relationship and rigour," (in eng), Fam Med Community

Health, vol. 7, no. 2, p. e000057, 2019, doi: 10.1136/fmch-

2018-000057.

[82] C. Willig, Introducing Qualitative Research in

Psychology. McGraw-Hill Education., 2013.

[83] UK Data Service. "Anonymising qualitative data."

https://ukdataservice.ac.uk/learning- hub/research-data-

management/anonymisation/anonymising-qualitative-

data/ (accessed 30-12,2024).

[84] M. Naeem, W. Ozuem, K. Howell, and S. Ranfagni, "A

Step-by-Step Process of Thematic Analysis to Develop a

Conceptual Model in Qualitative Research," International

Journal of Qualitative Methods, vol. 22, p.

16094069231205789, 2023/10/01 2023, doi:

10.1177/16094069231205789.

[85] W. Xu and K. Zammit, "Applying Thematic Analysis to

Education: A Hybrid Approach to Interpreting Data in

Practitioner Research," International Journal of

Qualitative Methods, vol. 19, p. 1609406920918810,

2020/01/01 2020, doi: 10.1177/1609406920918810.

[86] C. Davidson, "Transcription: Imperatives for Qualitative

Research," International Journal of Qualitative Methods,

vol. 8, no. 2, pp. 35-52, 2009/06/01 2009, doi:

10.1177/160940690900800206.

[87] L. S. Nowell, J. M. Norris, D. E. White, and N. J. Moules,

"Thematic Analysis: Striving to Meet the Trustworthiness

Criteria," International Journal of Qualitative Methods,

vol. 16, no. 1, p.1609406917733847, 2017/12/01 2017, doi:

10.1177/1609406917733847.

[88] IMO. "Convention on the International Regulations for

Preventing Collisions at Sea, 1972 (COLREGs)." IMO.

https://www.imo.org/en/About/Conventions/Pages/COL

REG.aspx (accessed 2025, IMO).

[89] IMO. "International Convention on Standards of

Training, Certification and Watchkeeping for Seafarers,

1978." IMO.

https://www.imo.org/en/OurWork/HumanElement/Page

s/STCW- Convention.aspx (accessed 08-03, 2025).

[90] National Transport Safety Board. "Contact of

Containership Dali with Francis Scott Key Bridge and

Subsequent Bridge Collapse." National Transport Safety

Board.

https://www.ntsb.gov/investigations/Pages/DCA24MM0

31.aspx (accessed 08-03, 2025).

[91] R. Sadeghi and F. Goerlandt, "A proposed validation

framework for the system theoretic process analysis

(STPA) technique," Safety Science, vol. 162, p. 106080,

2023/06/01/ 2023, doi:

https://doi.org/10.1016/j.ssci.2023.106080.

[92] European Maritime Safety Agency. "Risk Based

Assessment Tool (RBAT)." European Maritime Safety

Agency. https://www.emsa.europa.eu/mass/rbat.html

(accessed 08-03, 2025).

[93] N. El-Sheimy and A. Youssef, "Inertial sensors

technologies for navigation applications: state of the art

and future trends," Satellite Navigation, vol. 1, no. 1, p. 2,

2020/01/20 2020, doi: 10.1186/s43020-019-0001-5.

[94] F. Lázaro, R. Raulefs, W. Wang, F. Clazzer, and S. Plass,

"VHF Data Exchange System (VDES): an enabling

technology for maritime communications," CEAS Space

Journal, vol. 11, no. 1, pp. 55-63, 2019/03/01 2019, doi:

10.1007/s12567-018-0214-8.

[95] IMO. "Maritime Single Window." IMO.

https://www.imo.org/en/OurWork/Facilitation/Pages/M

aritimeSingleWindow-default.aspx (accessed 08-03,

2025).

[96] UK Civil Aviation Authority. "Reporting a cyber security

incident. Information and reporting procedures in the

event of a cyber security incident." CAA.

https://www.caa.co.uk/commercial- industry/cyber-

security/reporting-a-cyber-security-incident/ (accessed

20/02, 2025).

[97] Gov.UK. "Guidance - Where to Report a Cyber Incident."

Gov.UK. https://www.gov.uk/guidance/where-to-report-

a-cyber-incident (accessed 20/02, 2025).

[98] Gov.UK. "Guidance - Implementing the Network and

Information Systems Directive in the transport sector."

Gov.UK.

https://www.gov.uk/government/publications/impleme

nting-the- network-and-information-systems-directive-

in-the-transport-sector (accessed 20/02, 2025).

[99] M. Odéen, C. Sellberg, G. Praetorius, O. Lindwall, L.

Englund, and A. Andersson, "Intelligent Learning

Systems for Simulator-Based Professional Training: A

Systematic Literature Review," in Integrating Emerging

Technologies into Education and Training: Proceedings

of the 2nd ETELT 2024 Workshop, Cham, T. E. Kim, M.

Milrad, and I. Remolar, Eds., 2025// 2025: Springer Nature

Switzerland, pp. 109-119.

[100] R. Thomas, Modern Law of Marine Insurance Volume

Five, The (Maritime and Transport Law Library). London:

Informa Law from Routledge, 2023, p. 372.

[101] H. Ringbom, Røsæg, E., & Solvang, T., Autonomous

Ships and the Law London: Routledge, 2020, p. 318.

[102] R. Bo, "From Automatism to Autonomy,"

Conversations: The Journal of Cavellian Studies, no. 10,

pp. 98-127, 2023, doi: 10.18192/cjcs.vi10.6615.

[103] S. Chiodo, "Human autonomy, technological

automation (and reverse)," AI & SOCIETY, vol. 37, no. 1,

pp. 39-48, 2022/03/01 2022, doi: 10.1007/s00146-021-01149-

[104] E. J. de Visser, R. Pak, and T. H. Shaw, "From

‘automation’ to ‘autonomy’: the importance of trust

repair in human–machine interaction," Ergonomics, vol.

61, no. 10, pp. 1409-1427, 2018/10/03 2018, doi:

10.1080/00140139.2018.1457725.

[105] E. Veitch and O. Andreas Alsos, "A systematic review

of human-AI interaction in autonomous ship systems,"

Safety Science, vol. 152, p. 105778, 2022/08/01/ 2022, doi:

https://doi.org/10.1016/j.ssci.2022.105778.

[106] E. Veitch, O. A. Alsos, T. Cheng, K. Senderud, and I. B.

Utne, "Human factor influences on supervisory control of

remotely operated and autonomous vessels," Ocean

Engineering, vol. 299, p. 117257, 2024/05/01/ 2024, doi:

https://doi.org/10.1016/j.oceaneng.2024.117257.

643

[107] C. Campos, M. Castells-Sanabra, and C. Borén,

"Maritime Stakeholders’ Insights on Maritime Education

and Training in the Autonomous Shipping Era. A Small-

Scale Study in Spain " Scientific Journal of Maritime

Research - Pomorstvo, vol. 39, pp. 80-91, 2025, doi:

https://10.31217/p.39.1.7.

[108] S.-Y. Yi, M. Jung, and S.-I. Lee, "Adapting to change:

International maritime education and training for future

seafarers – focusing on the comprehensive review of the

STCW convention and code of the 10th session of the IMO

HTW," Journal of International Maritime Safety,

Environmental Affairs, and Shipping, p. 2464486, 2025,

doi: 10.1080/25725084.2025.2464486.

[109] IMO. "Sub-Committee on Human Element, Training

and Watchkeeping, 11th session (HTW 11), 10 – 14

February 2025." IMO.

https://www.imo.org/en/MediaCentre/MeetingSummari

es/Pages/HTW-11th-session.aspx (accessed 24/02, 2025).

[110] A. Connor, "Methods for the investigation of work and

human errors in rail engineering contexts," 07/15 2016.

[111] J. Wang, W. Fang, H. Qiu, and Y. Wang, "The Impact of

Automation Failure on Unmanned Aircraft System

Operators’ Performance, Workload, and Trust in

Automation," Drones, vol. 9, no. 3, doi:

10.3390/drones9030165.

[112] LUMIVERO. "NVIVO Qualitative Analysis Software."

LUMIVERO. https://lumivero.com/products/nvivo/

(accessed 20/02, 2025).