A study reveals 63% of maritime professionals use AI daily, yet only 8% report mature regulations and oversight in their companies.
According to a study by Thetius and Marcura, 63% of maritime professionals participating in the research use artificial intelligence daily. However, the proportion of those who state that the rules and oversight processes regarding this technology in their companies have matured remains at 8%.
Despite the integration of artificial intelligence into daily business processes in the maritime sector, it has been revealed that companies are not progressing at the same pace in terms of trust, human oversight, and accountability for decisions.
The report titled 'Earning Trust: AI in Maritime,' prepared by Thetius in collaboration with Marcura, is based on a survey involving 60 participants. The research addressed not only the frequency of technology use but also the time employees spend verifying the results produced.
It saves time despite the control burden
55% of participants indicated that they spend at least one hour per week checking or correcting outputs generated by artificial intelligence. The proportion of those who dedicate five hours or more per week to this task was 19%.
This additional workload was defined in the report as the 'verification cost.' Nevertheless, 85% of participants reported that even when accounting for the time spent on checking and correcting, artificial intelligence saves them time.
Human responsibility is maintained in critical decisions
When asked about their trust in AI results, 57% of participants chose the midpoint of a five-point scale.
Seven out of ten participants stated that they would definitely check the AI's recommendation before implementing an important business decision. 80% expressed that regardless of how advanced the technology becomes, the responsibility for a commercial decision should rest with a specific individual.
The proportion of those who stated that their company’s data is fully ready for the use of artificial intelligence remained below 10%. 47% of participants indicated that they would spend significantly less time verifying results that have proven reliability through experience.
Transition from recommendation systems to operational systems
The research also examined the transition from tools that only produce responses and suggestions to artificial intelligence systems capable of performing tasks.
72% of participants stated that their organizations are using, testing, or planning to use such systems. In contrast, the proportion of those who said that the relevant management and oversight processes are ready remained at 15%.
Marcura highlighted the importance of distinguishing between which operations can be performed independently by artificial intelligence and which require human oversight in its communication regarding the research.
Information is scattered across different systems
Janani Yagnamurthy, Senior Vice President responsible for Product and Strategic Growth at Marcura, stated that the fundamental issue in maritime is not a lack of data.
According to Yagnamurthy, the information needed to make decisions is scattered among contracts, emails, documents, various systems, and employees' experiences. For artificial intelligence to gain trust, it needs to demonstrate the information it relies on, be tested against real outcomes, and learn from human corrections.
Contributing to the sharing of experience
75% of participants reported that artificial intelligence helps less experienced employees gain effectiveness more quickly.
The report cited the example of CP Optimiser, developed by Marcura and Tomini Chartering. This tool compares charter agreements with the company's accumulated knowledge, helping to identify clauses and risks that experienced employees might notice.
Thomas Stjernholm, Commercial General Manager of Tomini Chartering, stated that in the process, humans provide the information, artificial intelligence associates this information with the relevant contract clause, and the final decision is again made by a human.
The report recommended that companies measure the entire process carried out jointly by humans and artificial intelligence, separate the necessary controls from repetitions arising from uncertainty, and clearly define decision-making responsibility.
Source: SeaNews Türkiye
