Sailing Through Legal Fog: How Regulation Lag is Stalling Maritime Automation

There’s no denying that AI systems, computer vision and automation are reshaping shipping operations. Safety at sea – not to mention voyage efficiency and energy consumption – is being optimized thanks to cutting-edge data collection and analysis that generate real-time insights for vessels and ports. Consequently, shipowners are increasingly equipping their vessels with advanced AI and computer-vision watchkeeping systems that support human decision-making.

At the same time, the pace of innovation is accelerating through the expert application of relevant technology, early adopters, pilot projects and a shared focus on overcoming operational challenges.

But to truly gain a foothold, innovation needs the support of effective regulatory and legal systems – which, unfortunately in today’s world, have yet to catch up to changing realities.

Legal framework out of step with technology development

Most of today’s maritime laws were penned for fully-manned ships, before digital and AI systems achieved robust navigational capabilities. Currently, the International Maritime Organization (IMO) does not yet have fully binding, dedicated rules governing AI-based decision support systems (DSS) or partially autonomous ships. 

But having identified these regulatory gaps, the IMO is currently working on creating and updating key frameworks such as the MASS (Maritime Autonomous Surface Ships) Code that can support the use of AI-enabled systems on the path to semi-autonomous shipping. Although existing safety, operational, and equipment regulations, specified in IMO frameworks such as SOLAS (International Convention for the Safety of Life at Sea) and the ISM (International Safety Management) Code, indirectly cover AI-powered systems installed on ships, AI and vessel automation is not directly covered by mandatory regulations. This leaves shipowners at the mercy of different interpretations of current laws by different national entities.

Mapping the challenges

Legal ambiguity regarding automation and AI-powered processes inadvertently create several conundrums. Here we focus on three: accountability grey areas; debates on seaworthiness; and standardization inconsistencies. 

Accountability Grey Areas

Legal ambiguity and the lack of regulations for smart ships creates “Grey Zones” regarding a very important aspect of commercial shipping: accountability. Today’s legal systems are yet to recognize the shift toward human-machine decision chains, which leads to serious liability issues. 

For example, in the scenario where damage occurs after a crew follows an AI system recommendation, it is unclear who is legally liable. This not only causes legal entanglements, but also leads to insurance claim complexities, as insurance frameworks are also slow to align themselves with AI or automated decision-making in maritime situations. 

Seaworthiness

While fully autonomous shipping has yet to become a feasible widespread reality, semi-autonomous shipping – a hybrid mix of machine and human tasks and commands – is well underway. But from a legal perspective, this presents a huge challenge that has yet to be fully addressed.

According to maritime law, in order to legally set sail a ship must be deemed “seaworthy” in terms of trained personnel and appropriate equipment on board. Yet as more ships adopt semi-autonomous practices, the question arises: what happens if the on-board AI system partially assumes navigational tasks, replacing humans? What happens if there are less personnel on board, deeming the ship semi-autonomous? Does this constitute a breach of “seaworthy” carriage clauses?  

Despite ongoing progress, today there is no binding international legal definition of seaworthiness that addresses semi-autonomous ships. Therefore, the courts have yet to set a precedent regarding the legality of semi-autonomous shipping, which is effectively stopping shipowners from adapting vessels to semi-autonomy. 

Standardtization Inconsistency

Smart ships are defined by high reliance on automation capabilities. However, classificiation societies do not have a unified view on what automated shipping is, applying different criteria in their analysis, assessments, and guidelines. Not all class standards are the same, nor are they legally binding. Similarly, flag states are often inconsistent in the way they view, review, and accept different aspects of ship autonomy. 

The IMO’s Maritime Safety Committee (MSC) has made attempts to create a regulatory framework for automated and autonomous maritime technologies, and the MASS Code is an ambitious effort to promote such advancements. According to the IMO’s timeline, the code should be finalized by May 2026. However, it is currently still defined as non-mandatory, meaning that each maritime jurisdiction can adopt it as it sees fit – or not at all. 

This can lead to situations where a certain jurisdiction adopts MASS recommendations and regulation, leading to cutting-edge progress, while another does the opposite – leading to confusion and lack of uniformity that thwarts effective standardization efforts. Having said that, the MSC has recognized the need for a mandatory Code, with adoption planned for 2030 and entry into force in 2032.

The Case for Phased Automation with Human-in-the-Loop

Amid this foggy legal reality, a middle path of phased automation makes most sense. This approach, which is advocated by Orca AI and other key technology players, is based on a “human-in-the-loop” concept that ensures crew members stay in full control, making all critical decisions regardless of their importance, while AI and computer vision systems provide robust support based on data analysis, insights, and recommendations. 

This approach allows automation to execute routine tasks such as navigation watchkeeping and  collision avoidance, with strict monitoring by crew who can override the system at any time. However, the impact of advanced AI automation – even with constant human monitoring – thrusts the “human in the loop” approach into the same muddy legal waters of full autonomy. Potential liability issues are still holding back the broad implementation this  innovative approach deserves. 

Sailing Forward: Bridging the Legal Gap 

The current under-adoption of robust AI systems is due to ship owners’ concerns that their vessels may be compromised in certain situations. Fleets that operate globally, with vessels registered in multiple flag states, tend to rely on operations and equipment governed by clear regulations. 

In situations where every flag can interpret the law on AI operation as they choose, many shipowners may hesitate to move forward with cutting-edge automation, even if it enhances situational awareness, fuel savings, and sustainability. There is therefore a pressing need to bridge the legal gap. Progress can be achieved through smart collaboration, including: 

• The IMO and International Association of Classification Societies (IACS) could join forces to outline clear standards for AI decision support, which, unlike full autonomy, is already feasible and making a big impact industry-wide.
  
• Flag states could kick-start national legal “sandboxes” that would pave the way for product pilots and solution testing in controlled environments. Similar mechanisms were deployed by many countries regarding autonomous and connected vehicles in the automotive sector.. 
• Finally, a cross-sector task force comprising insurers and legal bodies could tackle the issue of maritime liability head-on, in order to create effective models that reduce grey areas and clarify the legal basis of AI-assisted operations.

In conclusion, leveraging the benefits of AI and automation is critical for the future competitiveness of the maritime industry. International stakeholders must prioritize the development of clear digital regulations that cover the intricacies of AI-based decision-making and seaworthiness in semi-autonomous vessels. Increasing safety and efficiency at sea depends on it.