A new International Code of Safety for Maritime Autonomous Surface Ships
By Shani Friedman
Published on 15 June 2026
Introduction
In May 2026, the International Maritime Organisation (IMO) adopted a new International Code of Safety for Maritime Autonomous Surface Ships (MASS Code, or the Code). The new legal framework, which will be in effect from 1 July 2026, is a ground-breaking achievement—the culmination of nearly a decade of negotiations, adding another regulatory layer to the International Convention for the Safety of Life at Sea (SOLAS). The Code is currently not mandatory, although the IMO expressed its hope that it will become mandatory in the future.
While the final text is not publicly available, there is an advanced draft text that we can address.
This analysis aims to provide an initial assessment of the new legal framework and identify areas for further consideration.
The MASS Code
Although non-mandatory, the MASS Code is the first global framework that addresses and regulates Maritime Autonomous Surface Ships (MASS). The Code applies to ships that are either remotely operated or autonomous, even if there are crew onboard. The Code defines autonomous ships as ships whose equipment or systems were designed and proved to be controlled without human assistance or interaction (Proposed base text for the MASS Working Group, MSC 111/5/3, Preamble and Chapter 4, Articles 4.4 and 4.23), which suggests AI control. It is not enough to have high degree of automation; the main defining element of autonomous ships is in replacing (the functions of) onboard crew with automated or remote control. Similar to SOLAS, the new Code will not apply to warships (see SOLAS, Chapter I, Regulation 3; Proposed base text for the MASS Working Group, MSC 111/5/3, Chapter 2).
The Code sets out standards for safety, security and environmental protection of remotely controlled or autonomous ships, similar to those applied to regular ships. The IMO addresses MASS as a ‘goal-based’ framework, meaning it defines requirements and expected performances (Proposed base text for the MASS Working Group, MSC 111/5/3, e.g., Chapter 1, Part III).
While the MASS Code applies existing rules to remotely operated and autonomous ships, it also introduces new requirements for the design, approval and operation of these ships, including in key areas such as connectivity—meaning capabilities that allow MASS systems, Remote Operations Centres (ROCs), and other stakeholders to connect to a communication network (Proposed base text for the MASS Working Group, MSC 111/5/3, Preamble and Chapter 4, Article 4.7)—navigation, remote operations, and search and rescue. It requires risk assessment, robust system design, cybersecurity and the integration of ROCs.
With this new framework, the IMO tests its ability to adapt to and regulate new technologies. The organisation maintains that it allows for innovation (both technologically and legally), while ensuring safety and accountability, stressing that the human element remains at the core of maritime operations.
Issues for further consideration
Although the MASS Code is innovative and a remarkable achievement, certain issues pose concerns.
The first issue for consideration is the ships’ degree of autonomy in practice for the application of the Code. As mentioned, the Code applies and regulates ships that are either autonomous—i.e., at least some functions are performed without human intervention—or operated remotely. However, the Code also requires human oversight and control (Proposed base text for the MASS Working Group, MSC 111/5/3, Article 8.7) and human-centered system design (Article 9.2). This requirement seems to contradict the definition of ‘autonomous ship’ (Articles 4.4 and 4.23). This contradiction is acknowledged in the Code (see for example Article 8.7.2) but without further elaboration. Thus, it is questionable whether there can be a completely autonomous function in practice as defined in the Code. This question is strengthened by the Code’s requirement for a minimum safe manning of personnel onboard in accordance with the IMO’s past resolution (MSC 111/5/3, Articles 5.9, 13.2; IMO resolution A.1047(27)).
The second issue is that it is unclear whether remote operation (through ROCs) can be autonomous (meaning the systems onboard and ashore communicate with each other without human interaction) or if it must be operated by humans (see e.g., Proposed base text for the MASS Working Group, MSC 111/5/3, Chapter 4, Articles 4.33, 4.36, and 4.43). In the former scenario, the human element would be removed from the core of maritime operations, contrary to the IMO’s claim above. In the latter, the ship might not be considered ‘autonomous’ as envisioned by the Code. Thus, several principles under the Code are not applicable in practice. This last point is strengthened by the fact that the Code essentially requires consideration of ROCs in MASS and MASS functions, whether it is autonomous or remotely operated (see Proposed base text for the MASS Working Group, MSC 111/5/3, e.g., Articles 5.1.3, 5.2.2, 7.1, 7.8).
A third issue for consideration is the relationship between the MASS Code and the legal regime of the law of the sea. The Code was negotiated and adopted under the Maritime Safety Committee (MSC) of the IMO. This is a technical body, and its main competences are concerned with aiding navigation and safety. However, the Code specifically acknowledges the need to conform to the relevant rules of international law, including the United Nations Convention on the Law of the Sea (UNCLOS) (MSC 111/5/3, Preamble). There are two points in relation to which conformity with UNCLOS is at stake.
First, the draft resolution urges coastal states to allow MASS ships to enter their territorial waters, provided that these states are convinced that these ships fully comply with the Code (MSC 111/5/3, draft MSC resolution, p. 4). However, the Code does not address possible risks of MASS systems that could affect other relevant provisions of UNCLOS, such as the right of innocent passage in territorial waters under Article 17 of UNCLOS. Although mentioned in the Code, there is no in-depth consideration of cybersecurity risks which may affect the right to innocent passage. The Code generally refers to the risks to the MASS systems themselves rather than risks that can be caused by the MASS systems (MSC 111/5/3, e.g., Articles 7.8, 9.7). For example, issues such as unintended information gathering by MASS systems (onboard and ashore) might challenge the definition of innocent passage under Article 19 of UNCLOS, which in turn might hinder ships right to enter coastal states’ territorial waters. Unintended information gathering might threaten the security of the coastal state and in such cases, passage would not be considered innocent (Articles 19(1) and 19(2)(c) of UNCLOS). Furthermore, as mentioned, the Code does not elaborate on possible cybersecurity risks posed by autonomous MASS systems. MASS systems might also unintentionally interfere with a coastal state’s communication systems, which would not be considered innocent passage (Article 19(2)(k) of UNCLOS).
Second, the Code requires MASS operations to be ‘environmentally sound’ (MSC 111/5/3, Article 1.3.3), and that environmental protection be adequately assessed before approving MASS ships (Articles 6.1.1, 6.2). Expert personnel must carry out a risk assessment, including with respect to threats to the marine environment (Article 4.19, Chapter 7). However, again, there is no consideration how this Code could affect other relevant rules of international law, and specifically UNCLOS, despite the Code’s acknowledgement of their importance and connection to MASS. For example, UNCLOS regards pollution of the marine environment as an act by humans (Article 1(4) of UNCLOS). This raises the question of how pollution caused by autonomous ships will be regulated, if at all. While states must prevent, reduce and control pollution of the marine environment resulting from the use of technologies (Article 196 of UNCLOS), if there is no human intervention (as envisioned in the Code) and pollution is caused directly by technology, then this obligation might not apply to pollution caused by MASS ships as it could possibly not be considered ‘pollution’ at all. On the other hand, states still have the obligation to prevent pollution from vessels in general (Article 211 of UNCLOS), which may create a conflict not only between UNCLOS and the Code but between different provisions of UNCLOS. Preserving the human element in the Code at the expense of the degree of autonomy—as discussed above—might help solve this issue, despite the conceptual problems mentioned. Furthermore, in areas beyond national jurisdiction, the Agreement on Marine Biological Diversity of Areas Beyond National Jurisdiction (BBNJ Agreement), which entered into force on 17 January 2026, requires states to conduct environmental impact assessment (EIA)—a process to identify and evaluate potential impacts of an activity (Article 1(7), 28 of the BBNJ Agreement)—with respect to MASS. This may require additional monitoring beyond what is required in the Code.
Concluding remarks
The IMO’s MASS Code is an innovative and ground-braking achievement, introducing autonomous systems in shipping. The Code includes existing rules that apply also to conventional ships but also introduces new regulations for ships whose systems and equipment can operate without human intervention or are operated remotely. As the IMO has expressed its intention for the Code to become mandatory, further consideration must first be given to issues that may affect or be affected by MASS, most notably environmental protection, cyber risks, and the degree of autonomy that ships have in practice.
Dr. Shani Friedman, Senior Lecturer, Department of International Relations, School of Political Sciences, University of Haifa; Postdoctoral Fellow, Leonard Davis Institute for International Relations, the Hebrew University of Jerusalem. Her fields of expertise are international relations, international law, specifically the law of the sea, and international institutions. Twitter | LinkedIn