Lights and shadows of the application of technologies of artificial intelligence to the maritime and maritime sectors insurance companies have been examined in recent days by The International Propeller Club - Port of Genoa which organized a convivial meeting on the theme "Artificial intelligence and Marine and transport insurance: rules, risks and opportunities in the digital age".

Explaining that in the insurance sector, AI is already being used to make operations such as the claims management, customer service or personalization of the offers, Luca Bagnato of Prima Assicurazioni spoke both of the opportunities and risks associated with the use of artificial intelligence, the latter focusing on mainly on the reliability of data, on the protection of privacy and the proper use of sensitive information. At In this regard, he mentioned the European regulation on intelligence (AI Act), which classifies the different types of AI according to the risk (minimal, limited, high, unacceptable) and regulates its use.

Matteo Turci of the Turci Law Firm underlined how Artificial intelligence is changing the insurance, especially in the maritime and transport fields. In In particular, one of the most interesting areas is the use of to assess insurance risks. Compared to being AI can analyze a huge amount of data in a short time, making it possible to also ensure realities such as start-ups or businesses without a history Insurance. However, this advantage also comes with risks: often it is not possible to understand how AI arrives at its decisions, because it works as a sort of "black box" (black box). This - Turci pointed out - can create problems and for those who apply for a policy and are refused without reasons and for the insurer, who could accept risks too much high without realizing it. In addition, if the data used to training AI are manipulated, the result can be deliberately altered to take advantage of it.

Turchi also addressed the issue of self-driving vehicles, such as ships that travel without human manning thanks to sensors and artificial intelligence. In these cases - he observed - the rules are not yet ready to handle situations where there is no a recognizable human commander. In the event of an accident it becomes difficult to establish what happened, who is responsible, which insurance company should intervene and who should pay. Things get even more complicated when different artificial intelligences interact with each other, perhaps coming from by different and not always compatible companies, as happens in some automated ports. The presence of so many technological players makes it possible to It is difficult to understand where the problem lies and who has to answer for it.

Turci spoke about how AI and sensors can be used to automatically manage claims. For example, installing sensors in containers or in the holds of ships is possible detect any damage while the load is still in travel. In some cases, AI may start immediately reimbursement, even before the ship arrives at its destination. This allows for very fast handling, but can lead to mistakes: perhaps a refund is authorized for damage that does not exist, or you decide to throw away goods that are actually still good. To reduce this risk, one possible solution is use these systems only for minor damage or provide for always a final human verification.

According to Davide Pignone of Siat, today, however, the risk is not so much technological, but strategic: those who He does not start experimenting with artificial intelligence - he explained - risks being overtaken by those who know how to use it well. For For this reason, he recommended starting by mapping the processes to understand where AI can be useful, and to start small pilot projects to test it in a concrete way. Has also highlighted the importance of both training teams and managers, so that the whole organization can acquire a good knowledge of technology, and to integrate intelligence with the business, so that it becomes a natural part of the and advantageous of daily work.