Nowadays, the maritime industry, like other industries, is incorporating Machine Learning (ML) and Artificial Intelligence (AI) approaches in their applications. Since the rise of Maritime Autonomous Surface Ships (MASS) is on the horizon, such intelligent algorithms would replace conventional ship navigation with a higher level of autonomy. In other words, a digital navigator can be developed based on the data obtained from the human navigator's actions when controlling vessels. To ensure the prosperity of these vessels, the trustworthiness of such navigation actions must be guaranteed. Generally, the trustworthiness of any AI-based application can be studied from two primary levels: software and hardware. The software algorithms of trustworthy digital navigators should be Explainable, Fair, and Responsible. Besides, two concepts of Resilience and Availability must be confirmed for the hardware used for their development. Although the trustworthiness of the AI-based application from the software level is mainly focused on the previous research study, the trustworthiness of the hardware level should not be neglected. This preliminary study looks into ship systems used in such applications and then focuses on the digital navigator's trustworthiness at a hardware level. It identifies the most appropriate key performance indicators for studying this topic, and proper approaches to investigate them are summarized from the literature.

Source at https://onepetro.org/ISOPEIOPEC/proceedings/ISOPE24/All-ISOPE24/ISOPE-I-24-525/546284.