AI is going to be a game changer in the shipping industry – how will it affect how your firm will operate?

1. Introduction

Shipping industry is currently experiencing a convergence of challenges attributed to the side effects of the climate change and global supply chain disruptions. Decarbonization of the sea transportation, which forms part of the global climate change mitigation measures, will be coming at a significant cost of replacing the abundant fossil fuel energy and therefore shipping market participants need to explore ways to mitigate such cost that poses a direct threat to their traditional business model.

At the same time, many technologies, such as the Internet of Things (IoT), Machine Learning (ML), Artificial Intelligence (AI) and Blockchain, have been emerged and could potentially change this industry and disrupt traditional business models. Major changes are expected to occur both in the shipping design and operation (including alternative green fuels) and in the digitalization and collaboration of all parties of the shipping industry (dematerialization of all documentation, blockchain, digital twins, etc). This effectively leads us to the concept of the Digital Supply Chain - where all physical events in the real world should have a digital correspondent in the digital world, allowing more transparency and collaboration across the supply chain and shipping networks.

According to MIT, digital technology is accelerating the pace of business and transforming supply chains, which could in turn lead up to a 50% reduction in process costs and a 20% revenue increase. One of the main challenges is the inter-organizational collaboration changes during digital transformations. As an example, CMA-CGM and Google just announced an AI partnership agreement, to “actively seek to optimize vessel routes, container handling, and inventory management to ensure efficient and timely delivery of goods while minimizing costs and carbon footprints”.

At the same time, KPMG mentions “Generative AI in operations” in the top of the “Smart Supply Chain” (Supply chain trends 2024: The digital shake-up - KPMG Global) and Roland Berger, “anticipate that AI will reshape shipping over the coming years by optimizing flows, asset & network management, and profitability” (annexes).

Smart Shipping refers to the integration of technology and data in maritime operations. It encompasses various aspects, including autonomous vessels, data-driven decision-making, and onboard optimization. By leveraging digitalization, smart ships enhance safety, efficiency, and sustainability. Key benefits of Smart Shipping include transforming data into actionable insights. Ship owners and operators can optimize performance by accurately profiling ship behaviour and making informed decisions. Ultimately, Smart Shipping contributes to safer and more competitive maritime transportation.

According to Wartsila, the main factors that lead to the need to explore the possibilities of smart shipping are:

• The challenges related to the crew and the possibility of future shortage of human resources.
• Security.
• Human error.
• Climate change.

Smart shipping could serve as part of the solution of this difficult equation.

2. Key Industry Challenges

We have further elaborated on Wartsila’s list and have come up with the following list challenges and therefore areas of potential change/innovation that the maritime transportation industry needs to address/explore:

As Peter Drucker said “Innovations do not create change. Innovations succeed by exploiting change, not attempting to force it’’. Smart shipping can help address all the above factors. The automations that will result from the proper processing, interpretation and use of data will lead to remote control and diagnostic capability, which may ultimately lead to an automated ship. However, smart shipping and in general automations are and will continue to be very much associated with the management of data that is coming in in large volumes, high velocity and big variety. The maritime transportation industry is therefore called to manage this data that increases exponentially and utilize it in the day-to-day decision-making process. This is exactly where we see technology and in particular AI playing a vital role.

Big data analysis examines substantial amounts of data with the goal of extracting correlations and patterns (machine learning), as a better and deeper understanding of how specific things can evolve or change. At the same time, it can help companies make decisions that will be guided by these conclusions. The use of big data can have an exceptionally large and positive impact, offering to address those challenges. Big data is the “fuel” of AI and particularly of the deep (machine) learning process that enables computers simulating the human intelligence and its problem-solving capabilities.

There are several types of data in shipping:

• Navigation data: It includes data about the ship's position, speed, course. They are usually collected using technologies such as GPS, AIS, Radar, ECDIS.
• Ship performance data: Includes data related to fuel consumption, engine performance, speed, and other data related to ship operation.
• Weather data: Includes measurements such as wind speed and direction, wave height, temperature, rain depth, and other data that can affect the ship's operation. Main sources of weather data collection are meteorological stations and satellites.
• Cargo and freight data: This includes data about the type, weight and volume of the cargo being transported, as well as information about its condition and handling requirements, and the freight and pricing costs.
• Port Data: Includes data related to port operations such as ship movement in port and cargo handling. Usually, this data is collected by port authorities.
• Regulatory data: Includes data related to compliance with shipping-related regulations, such as exhaust emissions.

With so many types and sources of data, there are great opportunities to utilize them for the benefit of shipping. It is also evident that this utilization is multidimensional, with applications concerning the ship and its better performance, which in turn leads to a more environmentally friendly vessel operation. Security and autonomous ships are two more examples, with the latter being potentially a way forward in addressing the crew scarcity issues of the industry.

Additionally, the international trade is primarily conducted based on paper documentation. According to Mckinsey there could be a cost saving of $6.5 billion just by digitalizing the Bills of Lading. It is becoming apparent that here is still much to be done to digitalize all different forms of shipping documents such as the certificate of origin, the customs value declaration and the commercial invoice, as other examples. ?AI could facilitate a better management of all the information and data flows between shippers, carriers, shipping agents, forwarders, customs, terminals, ports and receivers. In the meantime, DCSA is working in defining the metadata to be used, in the new API standards (for track and trace, just in time port calls, etc.) for the container shipping business and has proclaimed a goal to fully implement eBL technology worldwide by 2030 (100% eBL by 2030 initiative).

3. AI (technology) tools/facilitators

I.??? Efficiency Improvement

• AI augmented weather routing solutions: one of the ways that can lead to a reduction in carbon dioxide and greenhouse gas emissions is weather routing. This optimization can analyze weather data from various sources. All this data can be collected from weather stations, satellites, gauges and sensors on ships.
• Use of data for machine learning aiming to route/speed optimization: take any voyage from port A to port B. The vessel could have one or more of several objectives: a) lowest fuel consumption and emissions, b) CII optimization or c) time charter equivalent maximization. Regardless of the goal one thing is true for every voyage, there is a single best route and speed plan to achieve it. But it is a complex formula to work out as several variables need to be considered. The impact of weather and oceanic conditions on fuel consumption varies with the ship type. The geometry, draft equipment and fouling state also make a difference. In other words, every vessel is unique and disregarding that simple fact results in burning up more fuel (different sources argue for 5%-10% fuel saving headroom). AI can be used by a company to develop a unique model for each of its fleet vessels and through this model work out the impact of winds, swell draft and currents in the ocean at any point in time and with high level of accuracy.
• Preventive maintenance: collecting data from machinery with AI informed algorithm and predict areas of potential failures.
• Facilitating better integration with the wider logistical chain: ship to port communication, just-in-time arrival and reducing congestion in ports.

II.?????????????????? Cyber security risk management

• Using AI to combat cyber threats: use AI and machine learning to identify abnormal behaviours that usually fraudsters adopt. This allows a company to score based on risk transactions, logins and sessions. This kind of information can help differentiate legitimate from fraudulent acts and lead to more informed decisions on the security steps required before further processing such transactions and session or accepting logins.

III.?????????????????? Navigational automations

• Collision avoidance: Advanced accident-avoidance systems use radar systems, sonar, AIS, meteorological data and historical data from past collision incidents and ship movements to provide a comprehensive picture and to be able to identify areas at risk, or even identify patterns that lead to collisions. Artificial intelligence also further enhances this capability through algorithms that continuously adapt to detect any anomaly or deviations of the ship from its normal behavior.
• Gesture recognition for bridge navigation.
• Taking route and speed automation to the next level: It is then on the company sole discretion whether to have its crew using those AI generated route and speed plan manually or take the automation to the next level by connecting these AI tools with the vessel’s main engine.

?????????? IV.?????????????????? Autonomous vessels

According to the IMO, there are four degrees of autonomy:

1st degree: Dispatched with automated processes and decision support. Sailors are on board to operate and control onboard systems and functions. Some operations may be automated and sometimes unattended, but with sailors on board ready to take control. 2nd degree: Remote controlled ship with sailors on board. The ship controlled and operated from another location. Sailors are available on board to take control and operate the ship's systems and operations. 3rd degree: Remote controlled ship without sailors on board. The ship controlled and operated from another location. There are no sailors on board. 4th degree: Fully autonomous ship. The ship's operating system can make decisions and determine actions on its own.

???????? V.??? Developing the next generation human capital

It's becoming a common discussion in workplaces across industries the possibility (or even probability) that several work specializations could become obsolete in the face of advancing technology and in particular AI.

The influence of AI on our future work processes implies it could also affect the skills we need to hire or cultivate in our teams. If AI can handle routine and repetitive tasks, it frees up time for individuals to learn new skills, take on new responsibilities, and engage in personal development.

As the shipping and maritime industries adopt AI, we need to manage the challenges of technological progress while keeping human-centred approaches at the forefront. Although AI has great potential to streamline operations and improve hiring processes, it is crucial to implement it ethically and transparently to reduce biases and ensure fairness.

Investing in employee training and development is vital to prepare the workforce with the necessary skills to succeed in an AI-driven environment. By leveraging AI as an innovation tool and promoting a culture of continuous learning, the shipping industry can adapt to the evolving work landscape and remain competitive in the global market.

??????? VI.??? Demand forecasting

Advanced big data analytics can aid in gathering and examining extensive datasets, such as historical demand trends, vessel operations, and financial information. This opens significant opportunities for creating more precise demand forecasts, enabling shipping companies to make more informed fleet management decisions.

Armed with the insights and data provided by AI and big data analytics, shipping professionals can enhance demand forecasting and effectively drive business growth.

Below we have summarized the key aspects of five case studies highlighting the use of AI in the shipping industry:

Source: How can AI be used in the Shipping Industry [5 Case Studies] [2024] - DigitalDefynd

4. Two faces how Generative AI is helping Shipping

4.1 Startups using AI to reinvent shipping industry

AI startups are revolutionizing the shipping industry through innovative solutions that enhance efficiency, safety, and profitability. We describe some of these examples according with the previous chapter:

·?????? Ship Angel: transforming ocean freight by integrating AI automation, which streamlines operations and reduces the need for manual intervention, thereby cutting costs and improving accuracy in shipment tracking and management.

·?????? Bearing AI: focused on optimizing vessel performance and fuel consumption by utilizing AI-powered weather routing, which helps ships navigate more efficiently, reducing fuel consumption and emissions.

·?????? Captain AI: pioneering the development of fully autonomous shipping solutions, aiming to create the world's first safe autonomous vessels. This advancement promises to enhance maritime safety and operational efficiency by minimizing human error and enabling more precise navigation.

·?????? Commodity AI: automating shipment management for commodity trading operations, providing tools that offer greater control over logistics, thereby optimizing supply chain processes and reducing operational risks.

·?????? Breeze AI: offering a fully automated and digital insurance solution tailored specifically for freight forwarders and logistics companies. Their machine learning platform represents an industry-first, delivering customized insurance products that are seamlessly integrated into the logistics workflow, enhancing risk management and simplifying the insurance process.

These startups, along with others in the field, are collectively pushing the boundaries of what is possible in the shipping industry. They are leveraging AI to create more intelligent, efficient, and resilient logistics networks, which in turn are helping to meet the growing demands of global trade while addressing environmental and safety concerns. As AI technology continues to evolve, the impact of these innovations is likely to expand, further transforming the shipping landscape.

4.2 Startups and Scaleups using AI to accelerate their development

Generative AI tools are revolutionizing the way software companies (and large shipping companies, like Maersk) develop code and perform testing by automating and accelerating various aspects of the software development lifecycle. These tools are making a significant impact in several ways.

Generative AI tools can analyse patterns in existing code and generate new lines of code that are optimized for readability, efficiency, and error-free execution. They assist with code review by identifying potential issues, suggesting improvements, and helping maintain high-quality codebases. This capability ensures that the development process is not only faster but also results in more reliable and maintainable software.

In terms of automated testing, generative AI tools can generate test cases, perform automated testing, and even debug code. This significantly speeds up the testing process, allowing developers to identify and fix issues more quickly and efficiently. This automation reduces the time and effort traditionally required for thorough testing, thus accelerating the overall development cycle.

While no-code/low-code platforms enable rapid development with minimal coding, generative AI tools take this a step further by generating code that didn’t exist before, offering more customization and complexity. These tools can produce entire functions in response to natural language prompts, providing a level of flexibility and power beyond the predefined components of no-code/low-code platforms. This capability allows for the creation of more sophisticated and tailored software solutions.

Performance gains from using generative AI tools are well-documented. Studies have shown that tools like GitHub Copilot can lead to significant reductions in developer labour cost. For example, one study indicated up to 50% of time saved in code documentation and autocompletion, and 30-40% in repetitive coding tasks, unit test generation, debugging, and pair programming. Another study found that GitHub Copilot could lead to a 33-36% time reduction for coding-related tasks in a cloud-first software development lifecycle. These performance gains translate into faster project completion and more efficient use of developer resources.

Furthermore, research has also focused on the impact of tools like GitHub Copilot on developer productivity and happiness. The findings suggest that these tools can indeed improve the efficiency and satisfaction of developers, making their work more enjoyable and less tedious. Enhanced productivity and developer happiness are critical factors in maintaining a motivated and effective development team.

In summary, generative AI tools like GitHub Copilot and similar platforms are not only speeding up the code development and testing processes but also providing substantial performance gains. They offer a viable alternative to no-code/low-code platforms by enabling more complex and tailored software solutions. The studies conducted so far are promising, indicating that these tools can significantly enhance productivity and reduce the time spent on various development tasks.

5. Potential implementation barriers

AI adoption in shipping holds promise for emissions/cost reduction, improved operational efficiency, increased safety and security and better forecasting in the coming decades, but there remain significant practical, legal and regulatory challenges that need to be addressed before this technology can be widely implemented across fleets. Those challenges are:

I.??? Ship connectivity

Limited bandwidth and coverage issues have consistently hindered internet connectivity at sea. Yet, innovations that monitor and track fuel consumption, performance, and provide real-time weather updates all depend on stable and dependable bandwidth. The evolution of low earth orbit satellite communication services (such as the Space X-Starlink) is expected to partially address those coverage issues and is expected to facilitate better crew communication and welfare as well as significantly improve connectivity between the office and the vessel.

II.? Insurance coverage issues

Insurance policies need to be redefined to cover scenarios with accidents involving smart (autonomous) vessels. Terms like "duty of care'' and ''negligence'' need to be reconsidered in a future when machines will conduct duties that were previously conducted by human beings. Defining the root of causation will be tricky as the reference is always against “humans” and the question on how you support causation if something goes wrong remains unanswered.

III. Intellectual property – copyright issues

In section 2, we touched upon the different types and sources of data used in the shipping industry including navigation data, cargo and freight data, trading patterns, emissions data etc. All this data will be constantly utilized by AI tools to produce useful insights for the end users. Questions like:

i) who is the owner of this data,

ii) can it be copyrighted and

iii) can it be sold

Need to be properly addressed to facilitate wider implementation across the industry participants.

IV. Standardization gap

Flag/IMO standards on navigational performance assume that vessels are navigated by humans. All international standards and codes need to be redefined to adapt to a new era with less human intervention.

V. Government intervention

The final form of several regulatory acts aiming to provide a risk management framework in the use of AI (i.e. EU AI act) will significantly affect the level of adoption of the developed AI applications. Stringent rules are expected to significantly affect the implementation of those called ''high risk applications'' in shipping including but not limited to the autonomous navigation, crew health records, preventive maintenance of critical vessel equipment and cyber security systems

6. Conclusion

In summary, the shipping industry finds itself at a critical crossroad, facing an array of challenges mainly due to climate change and supply chain disruptions. However, the emergence of cutting-edge technologies—such as IoT, ML, AI, and blockchain—presents transformative opportunities. These innovations can reshape shipping practices, elevate digitalization efforts, and foster seamless collaboration across the entire supply chain. Notably, the concept of the Digital Supply Chain, which advocates for a digital counterpart to every physical event, promises substantial benefits, including cost savings and revenue growth. Real-world examples, like the AI partnership between CMA-CGM and Google, underscore the practical applications and advantages of these technological strides.

Smart shipping, a pivotal innovation, seamlessly integrates data and technology into maritime operations. By harnessing digital tools, smart shipping enhances safety, operational efficiency, and environmental sustainability. Key advantages include translating data into actionable insights, optimizing vessel performance, and potentially automating entire ship operations. Despite the inherent challenges—ranging from cybersecurity threats to crew shortages—smart shipping offers pragmatic solutions.

AI plays a central role in this transformative journey, offering advanced capabilities such as precise weather routing, predictive maintenance, fortified cybersecurity, and automated navigation. It also supports the development of autonomous vessels and ensures that human skills remain competitive in an AI-driven environment. Furthermore, AI-powered big data analytics enhance demand forecasting, enabling better fleet management and business growth.

Yet, significant barriers persist. Connectivity gaps, complex insurance considerations, intellectual property issues, and regulatory hurdles demand collective efforts from industry stakeholders and regulatory bodies.

Generative AI tools further accelerate technological progress within the shipping sector. By automating code generation, testing, and debugging, these tools not only boost efficiency and reduce developer workload but also empower the creation of sophisticated, customized software solutions—yielding a competitive edge in the dynamic market.

In essence, the fusion of AI and digital technologies within the maritime industry holds immense promise. It addresses current challenges while propelling future growth. As these technologies evolve, their impact will expand, reshaping the shipping landscape and fostering a more efficient, sustainable, and collaborative global supply chain.

7. Annexes