There are four main pillars for artificial intelligence (AI) that can be embedded in port and terminal operations.
- Predictive AI. Focuses on analyzing trends and patterns that can be used to forecast or detect anomalies using a wide range of statistical methods. It tends to reshape organizational expertise and expand the capabilities to understand factors that influence current and future port and terminal operations. The challenge is that predictive AI may blur the process by which the results were achieved and undermine the cognitive abilities of the organization as the expertise is outsourced to an automated party. For the port and its terminals, it provides expanded operations analytics, which includes market conditions, equipment monitoring, and performance metrics. Dashboards are common in the port industry.
- Generative AI. Often informed by predictive AI, generative AI is able to generate content, including reports, data, or documentation that allows the planning of operations. The work process becomes more focused on the curation and verification of the documentation, and less on its production. This can challenge the authorship of the documentation, including who it may be attributed to.
- Agentic AI. Involves forms of reasoning and tasks that allow an AI to act on behalf of its issuer in decision-making. The challenge with automated decisions and actions is accountability: who actually has authority and sovereignty over them? Terminal operating systems are usually able to support operations and plan accordingly, including allocating resources such as equipment, yard storage, and inland transportation.
- Embodied AI. Integration of AI with physical forms to perform the equivalent of physical labor or labor controlling physical flows. This allows tasks to be repeated safely and in compliance with established procedures. Automation allows for the delivery of terminal operations.